Conference Handbook and Delegate List 4-6 September 2012 The University of Edinburgh CONTENTS PAGE Welcome Message from the Chair of OR53 3 Conference Organising Committee 4 Stream Organisers 5 Sponsors, Exhibitors and Advertisers 6 Practical Information 13 OR54 Goes Mobile ” Guidebook App 18 Conference Programme at a Glance 19 Conference Timetable 23 Plenary Sessions 33 President Medal Presentations 37 Abstracts: Analytics 39 Community OR and Sustainable Support for Communities 41 Credit Risk Management 44 Forecasting, Data Mining & Computationally Intensive Methods 47 Government O.R. Service 53 Green Logistics 54 Health 59 Horizon Scanning & Future Analysis 67 Information Systems & Knowledge Management 69 Logistics and Supply Chain 71 Making an Impact 75 Metaheuristics 81 Operations-Design-Innovation Interface 83 Optimisation 86 OR & Strategy 91 OR Consultancy and Case Studies 96 1 O.R. in Education 100 Problem Structuring Methods 102 Project Management 105 Queue Modelling 108 Routing Applications & Transportation 110 Scheduling 114 Simulation 117 Stochastic Inventory Control 122 Stochastic Scheduling and Dynamic Allocation Problem 124 Speaking Authors Index 126 Delegate List 134 Walking Map of Arthur’s Seat 139 Session Room Floor Plans Inside Back Cover Conference Campus Plan Back Cover 2 Welcome Message from the Chair of OR54 A very warm welcome to the 54th conference of the OR Society and to the University of Edinburgh. The programme for the next three days includes more than 200 presentations by practitioners and academics covering a remarkable variety of topics. After great success at the last two conferences, the ‚Making an Impact‛ stream returns this year on Wednesday. Aimed at practitioners, but open to all, this stream consists of a range of talks and workshops designed to promote the exchange of ideas and experience relevant to practice. As always the technical programme is complemented by a social programme that is similarly varied. This year we have excursions to three local attractions that are below the radar of the average Edinburgh visitor: Glenkinchie Distillery, Lauriston Castle and the National Museum of Flight. Following tradition, the social programme also features the ever-popular quiz night on Tuesday and a gala dinner on Wednesday evening. I am particularly looking forward to the four plenary talks. Society President Geoff Royston will open the conference on Tuesday with his Presidential Address: O.R. for the real world: big questions from a small island. Also on Tuesday, John Hope of Ernst & Young will survey the emerging field of analytics and its relationship with OR. This session will provide an opportunity for discussion of the Society’s analytics initiative. On Wednesday, Jason Field of Dstl will explain the role of OR in supporting decision makers dealing with current day defence and security challenges. Finally on Thursday, Chris Dent of Durham University will discuss the application of O.R. to energy planning, focussing on the three challenges of data, optimisation and futurology. I am delighted to have such a strong and varied series of plenary talks at the conference. There will be one further plenary session on Wednesday afternoon in which the three finalists for the President’s Medal present their work. The Medal is awarded for the best practical application of OR. The applications addressed by this year’s finalists are modelling demand and performance in HMRC call centres, decision analysis in nuclear decommissioning and shaping the NATO plan for Afghanistan. I encourage everyone to attend this session and have their say. Conferences like this are only possible thanks to the hard work of a great many people. Firstly I give my sincere thanks to my colleagues on the Organising Committee „ Robert Raeside, Galina Andreeva, Kerem Akartunali, Amanda Tucker, Rosemary Byde and Hilary Wilkes „ for their hard work and support throughout the year. Next I thank the many stream organisers for inviting, persuading and cajoling authors to present their work at the conference. I am very grateful to the sponsors and exhibitors who have supported the conference despite the continuing difficult economic conditions. The aforementioned contributions have created an excellent programme of events which I hope everyone will enjoy. It only remains for me to thank all of you for participating in the conference and helping to make it a success. I hope OR54 will live long in the memory. Tom Archibald, University of Edinburgh 3 Conference Organising Committee As our OR54 Chair has said, conferences like this are only possible thanks to the hard work of a great many people, not least of which are our colleagues on the Organising Committee - Tom Archibald, Robert Raeside, Galina Andreeva, Kerem Akartunali, Amanda Tucker, Rosemary Byde and Hilary Wilkes „ thanks to them for their hard work in making this conference a success. The OR Society Conference Chair Tom Archibald University of Edinburgh T.Archibald@ed.ac.uk Programme Scheduler Kerem Akartunali University of Strathclyde Business School kerem.akartunali@strath.ac.uk Programme Co-ordinators and Keynote Editors Galina Andreeva University of Edinburgh galina.andreeva@ed.ac.uk Robert Raeside Edinburgh Napier University r.raeside@napier.ac.uk Social Events Organiser Amanda Tucker Dstl atucker@mail.dstl.gov.uk Sponsorship and Exhibition Organiser Rosemary Byde Royal Bank of Scotland Rosemary.Byde@rbs.co.uk Administrator Hilary Wilkes The Operational Research Society Hilary.wilkes@theorsociety.com 4 Stream Organisers Stream Name Organiser Analytics Jacqui Taylor Community O.R. and Sustainable Support Martha Vahl and Eliseo Vilalta-Perdomo for Communities Credit Risk Management Jonathan Crook Forecasting, Data Mining & Computationally intensive methods Fotios Petropolous and John Crocker Green Logistics Tolga Bektas Health Sonya Crowe Horizon Scanning and Future Analysis Amanda Tucker Information Systems & Knowledge Management Jo Smedley Logistics & Supply Chain Patrick Beullens Making an Impact Mark Roper, John Ranyard and Ruth Kaufman Metaheuristics Ender Ozcan and Andrew Parkes Operations-Design-Innovation Interface Shu-Jung Sunny Yang Optimisation Konstantinos Kaparis O.R. & Strategy Frances O’Brien and Martin Kunc O.R. Consultancy and Case Studies Kuangyi Liu and Sophie Carr O.R. in Education Louise Orpin Problem Structuring Methods Edward Terry Seagriff Project Management Gary Bell, Rosane Pagano and Jon Warwick Queue Modelling Navid Izady Routing Applications & Transportation Luc Muyldermans and Jamal Ouenniche Scheduling Djamila Ouelhadj Simulation Kathryn Hoad Stochastic Inventory Control Roberto Rossi Stochastic Scheduling & Dynamic Allocation Problems Didac Ruiz-Hernández The Operational Research Society would like to extend a warm thank you to all the stream organisers, who have successfully devoted a great deal of time and effort in putting in place an excellent mix of papers. Where applicable, the Society would also like to thank the employers of all those involved in organising the conference for providing them with the opportunity to contribute in this way. 5 Sponsors, Exhibitors and Advertisers We are very pleased to acknowledge the support of all our sponsors, exhibitors and advertisers. The following organisations have sponsored some aspect of the conference: Dstl, (Keynote handbook and Memory sticks) Edinburgh Napier Business School, (General sponsorship) Palgrave Macmillan, (Gala evening drinks reception) University of Edinburgh Business School, (Plenary) 2020 Consulting, (Gala evening drinks reception) whilst the following have exhibition stands or poster boards: Aimms Decision Lab 2020 Consulting NATCOR Palgrave Macmillan Palisade SIMUL8 Corporation Simulation Solutions University of Southampton Wiley-Blackwell Stream sponsorship has been provided by: Banxia Software Paragon Business Solutions SIMUL8 Corporation Inclusion of promotional material in delegate packs has been commissioned by: Banxia Software Elsevier GORS Simulation Solutions Taylor and Francis University of Southampton Please take a moment to learn more about all our supporters below 6 Dstl is a trading fund of the Ministry of Defence (MOD), delivering trusted and often confidential advice and solutions on defence-related science and technology that impact on the security of the UK. It’s demanding, exciting and rewarding work and each year Dstl recruits approximately 150 graduates to be part of it. Graduates receive a balanced programme of courses to develop technical and non-technical skills. Sponsorship is available to gain Chartered status and take further Qualifications. There are also opportunities for secondments, both in the UK and overseas. Details of Dstl’s current graduate vacancies can be found on our website at www.dstl.gov.uk/careers Edinburgh Napier University is one of Scotland's top universities for graduate employability. 93.2% of graduates are in work or further study within six months of leaving. This university is also proud winner of the Queen's Anniversary Prize for Higher and Further Education 2009, awarded for innovative housing construction for environmental benefit and quality of life. Palgrave Macmillan is proud to be the publisher of The Operational Research Society and sponsor of this year's drinks reception, where we are celebrating KMRP's first decade of publication and OR Insight's 25th anniversary year. We look forward to talking to you at our stand where we will have our complete O.R. portfolio on display, including copies of the newest OR Society journal, Health Systems. Take advantage of the great membership offers on both the Society journals and our related titles. The University of Edinburgh Business School enjoys a long tradition of teaching and research. The School offers undergraduate, postgraduate and executive education programmes in business and management and provides a platform for research, discussion and debate on a wide range of business issues. Part of the University of Edinburgh, one of the world’s top 20 universities with a rich heritage of delivering education for over 400 years, the Business School has an international student body typically representing more than 88 countries. The student body comprises 800 undergraduate students, 430 postgraduate students and 90 Doctoral students. 7 Recently the School relocated to a new building at Buccleuch Place, located at the heart of the University campus. This state-of-the-art space features eight lecture theatres, multiple syndicate rooms, an executive education suite, student study centre and online resource, café and significant flexible space for staff and students. The building has been designed and specified to meet the current and anticipated future needs of the School’s portfolio of activities and represents an investment of £17m by the University. The School’s teaching and research covers six main subject areas ” accounting & finance, entrepreneurship & innovation, management science & business economics, marketing, organisational studies and strategy & international business. The School is accredited by EQUIS and AMBA and consistently ranked in the Financial Times and Economist MBA Rankings ” reflecting not just a long history of business teaching but also the substantial experience of a faculty comprising more than 80 teaching staff. AIMMS is the Advanced Integrated Mathematical Modeling System, a software tool for developing and deploying optimisation models. AIMMS enables fast model formulation in a tree structure without having to learn a programming language, and links seamlessly to a range of mathematical solvers. AIMMS offers procedural execution and flexible modelling concepts such as rolling horizons, column generation, outer approximation for MINLP, multi-start for NLP, Stochastic Programming, Benders decomposition, Robust Optimization, etc. AIMMS offers an integrated GUI to visualise and interpret results. Data can be exchanged with many sources, from XML- and text-files to Excel and databases. Finally, models can be integrated with other applications via options such as the COM object, the Excel add-in and Web Services. www.AIMMS.com Decision Lab is a pioneering consultancy which combines commercial acumen and specialist modelling expertise to support strategic, tactical and operational decision making. Our team of problem structuring, optimisation, simulation and business modelling professionals aims to significantly improve the strength and resilience of your medium and long term business critical decisions. By providing an environment in which to explore options and test scenarios, Decision Lab can help reduce the risk associated with making strategic choices. Through our collaborative approach and use of sophisticated modelling techniques we enable clients to safely explore risk, quantify uncertainty, identify the optimal course of action and make better, more resilient evidence-based decisions. Specialties Optimisation, Agent Based Simulation, Discrete Event Simulation, System Dynamics, Problem Definition & Structuring, Operational Research, Business Modelling. 8 A new breed of specialist consulting firm Our consultants have in-depth knowledge of industry sectors that include financial services, travel & leisure, automotive and utilities with our primary areas of expertise in risk, marketing, digital, data management and technology. We add to that our talent for marshalling data from multiple sources, to arrive at a deeper understanding of customer behaviour. We love Big Data. Every day, 2.5 billion billion bytes of data are created. We love nothing more than making sense of this for our clients, using our exceptional analytical skills to separate value-creating insight from noise. You don’t get anywhere standing still. We believe in the power of new. Harnessing the potential of new customers, new services, new channels, new markets and new ways of working is the lifeblood of growth. Making change a truly positive and profitable force is the kind of challenge we thrive on. What could new do for you? 20:20 Jaywing will become 2020 Consulting soon - look out for changes to weare2020.com. Meanwhile, email information@weare2020.com or call Ben O’Brien on 08000 733393. NATCOR is a collaboration between ten universities to develop and deliver taught courses in Operational Research (O.R.) to PhD students. The initiative to develop such taught course provision came from the UK Engineering and Physical Sciences Research Council (EPSRC). It is part of a drive to deepen and broaden PhD studies in the UK across the mathematical sciences. NATCOR is funded by EPSRC for the first five years of its life (October 2006 to September 2011) and is supported by the OR Society which is a collaborating partner. Founded in 1984, Palisade is the world’s leading developer of risk and decision analysis software. Products including @RISK and the DecisionTools Suite integrate seamlessly with Microsoft Excel to bring techniques such as Monte Carlo simulation, optimisation and neural networking tools to your desktop in a familiar and intuitive environment. Palisade software is now on the syllabus at many top universities and business schools across Europe and North America - from operational research and MBA programmes through to engineering and accountancy 9 degrees. They have also been adopted by many of the world’s biggest companies, including 93 out of the Fortune 100 companies and most of the FTSE 100. 25 years of development and innovation means that performing quantitative risk and decision analysis has never been easier than it is with @RISK and the DecisionTools Suite. Find out how you and your university can benefit from using the software by downloading a free trial from our website, www.palisade.com. Alternatively, please email cmarlow@palisade.com to request case studies, white papers or to arrange a free web demonstration. SIMUL8 simulation software enhances the way businesses make and communicate decisions. Our software delivers multi-million pound savings for major organisations - including the US Government, GM, the NHS, and Coca-Cola - by allowing them to experiment with process improvement ideas in a risk free environment. SIMUL8 is also used to teach simulation in elite Universities across the world including the London School of Economics, Melbourne Business School, Singapore National University and Hong Kong City University. Simulation Solutions will be demonstrating Siemens Plant Simulation an object-oriented simulation tool that allows faster development of large models by supporting object oriented concepts including classes, methods and inheritance. Plant Simulation enables the simulation and optimisation of production systems and processes. Using Plant Simulation, you can optimise material flow, resource utilisation and logistics for all levels of plant planning from global production facilities, through local plants, to specific lines. CORMSIS (Centre for Operational Research, Management Sciences and Information Systems) Advanced mathematical and analytical modelling methods for better decisions CORMSIS at the University of Southampton is the largest and most comprehensive groups of OR/MS/IS researchers in the UK, with over 30 academics and over 60 research students. It covers the whole spectrum of current OR/MS/IS from theoretical mathematical developments to problem structuring and knowledge management, and has internationally-renowned expertise in the specific areas of risk, optimisation, finance and health. We have experience in consultancy, student placements and joint research. We focus our research on the practical application of these tools and techniques to solve ‘real’ business problems for ‘real’ organisations. We have collaborated with a wide range of external organisations, via our summer project scheme and other research initiatives. Our summer project scheme offers organisations the 10 opportunity to have one of our MSc students analyse a specific business issue, with structured academic support from leading academics. This scheme has been running for over 10 years and resulted in many £’s of savings and additional revenue for these organisations. For examples of this work see the brochure in your delegate pack or visit our website. Email: cormsis@soton.ac.uk Web: www.soton.ac.uk/cormsis At Wiley, our mission is to help teachers teach and students learn. We are committed to publishing and making available the best educational and training materials for all types of instructors: textbooks, electronic teaching aids and products to help you be an even better teacher. Wiley Higher Education serves undergraduate, graduate and advanced placement students, and lifelong learners. We publish educational materials in all media, notably through WileyPLUS, our integrated online suite of teaching and learning resources. Wiley-Blackwell publishes more than 75 journals in Business and Management, partnering with leading societies in the field. Our journals are available in thousands of institutions globally, and are among the most widely read, frequently cited, and top-ranked journals in the ISI's Business and Management categories. Please visit our stand at the conference, where we will be displaying our new and classic books and journals in business and management, and enjoy a 20% discount on all orders. Banxia Software Ltd develops and sells high quality decision support (DSS) and audience response/ classroom participation systems. We provide training in the use of our software and systems and also develop add-ons for other systems. As ‘Top Interwrite PRS Partner’ we have supported many universities throughout the UK in their use of Interwrite™ Response. Our products are used in the public, private, military, not-for-profit and academic sectors. Our offerings include Interwrite™ Response, Impact Explorer™, Decision Explorer® (for cognitive/ causal mapping) and Frontier Analyst® (for DEA assessments). Paragon, established in 1991, is the UK’s leading credit scoring software solution provider. We offer a comprehensive range of credit management and decision making tools. Our software includes Modeller, a fully integrated model development environment, for the building of risk management scorecards and decision systems. Paragon’s decision engine is a flexible system suited to application scoring, behaviour scoring, and business strategy implementation. 11 Our pioneering tools are used worldwide by major banks, finance houses, mail order companies, insurance services, and credit scoring consultants and benefits from our close relationship with leading academic and credit industry researchers. Our strategy is to provide solutions for our clients’ ever-changing business requirements. Elsevier Operational Research and Management Science Elsevier is the leading journals publisher in the field, working in partnership with the global academic community to publish 12 of 20 top-ranked journals in the ‘Operational Research and Management Science’ category in the Social Citation Index. Keep up to date with all our journals including the new journal Operations Research for Health Care at www.elsevier.com/decisionsciences or tweet with us @ORMS_Elsevier GORS ” the Government Operational Research Service has joined the Civil Service Analytical Fast Stream and is hoping to recruit 20-30 Operational Research Analysts in 2011/12. Further details can be found on the GORS website www.operational-research.gov.uk. The Taylor & Francis Group publishes more than 1,500 journals and around 1,800 new books each year, with a books backlist in excess of 20,000 specialist titles. We are providers of quality information and knowledge that enable our customers to perform their jobs efficiently, continue their education, and help contribute to the advancement of their chosen markets. Our customers are researchers, students, academics and, increasingly, professionals. 12 Practical Information Web address for The OR Society conference pages - www.theorsociety.com/OR54 Main Tel +44 (0)121 233 9300 Web address for the John McIntyre Conference Centre (JMCC) at Edinburgh First: http://www.edinburghfirst.co.uk/venues/john-mcintyre-conference-centre Main Tel. for Edinburgh First Reception +44 (0)131 667 1971 (Switchboard), +44 (0)131 651 2189 (Events Office). Conference Conference proceedings open with the Chair’s welcoming address at 09:30 on Tuesday 4 September 2012 and conclude at 15:00 on Thursday 6 September 2012. Registration Conference registration will take place in the John McIntyre Centre. Tues, Wed and Thur from 08:00. Accommodation Check-in time is from 14:00. Check-out time is 10:30 latest. Luggage can be stored either within the Edinburgh First Reception centre or the Events team will arrange for a separate room for this purpose. Bedrooms must be vacated by latest 10:30. Room keys are available from the main reception area of the Edinburgh First Reception from 14:00 on Monday 3 September 2012. All bedrooms are located in the Pollock Halls residential area, in Chancellors Court, a few minutes’ walk from Edinburgh First Centre and the JMCC building. Edinburgh First reception will be open 24 hours, 7 days a week and the contact telephone number is Tel: +44 (0)131 667 1971. All bedrooms have:Tea and coffee making facilities in each bedroom. Towels and basic toiletries in each bedroom. Ironing facilities are located in the pantry areas on each floor. Hairdryers are not provided in the bedrooms but can be obtained at the reception desk, if required. Wired Internet access (not wifi), is available in the bedrooms. The charges are as follows:1 Day = £5, 2 ” 3 Days = £10 or 4 ” 7 Days = £15.00. Contact Edinburgh First Reception for further information. (A data cable is in each of the bedrooms for guests to use, although you may wish to bring your own). You are advised to bring some cash for drinks and other sundry items as individual room accounts cannot be set up. 13 Meals and drinks (see also the Socials Events below) For residential delegates and those who have booked meals as a separate option: With the exception of the Gala dinner on Wednesday evening, all meals will be served in the JMCC restaurant on the ground floor. Tuesday evening: Centro Bar ‘cash’ bar will be open 19:30 ” Midnight. (1st floor JMCC) Wednesday evening: Delegates attending the Making an Impact Practitioner sessions are invited to enjoy a glass of wine in the JMCC Centro bar area at 18:00. All delegates are invited to attend a drinks reception which will be held in St. Leonards Hall from 19:15 ” 20:00. Gala Dinner: will be served in South Hall (a few minutes’ walk from St. Leonards Hall) from 20:00. Breakfast each day will be served from 07:00 ” 10:00. Dinner availability after 20:30 may be available in the JMCC bar on the ground floor until 21:30. Tuesday 4 Sept. Breakfast Lunch Dinner 07:00 ” 10:00 12:30 ” 13:30 18:30 ” 20:30 A cash bar will be open in the Centro Bar, first floor in JMCC 19:30 ” Midnight Wednesday 5 Sept Breakfast Lunch Drinks Reception Evening Gala Dinner 07:00 ”10:00 12:30 ” 13:15 19:15 ” 20:00 St. Leonards Hall 20:00 ” 22:00 South Hall. Please bring your ticket with you. A cash bar will be open in South Hall until Midnight ” (see Socials section below). Alternatively the JMCC bar on the ground floor of the JMCC building will be open until 11:00 (although this is not exclusively for OR Society use). Thursday 6 Sept Breakfast Lunch 07:00 ” 10:00 12:30 ” 13:30 Badges will be required to gain access to the JMCC restaurant at all times. Notes for Speakers / Stream Chairs For Speakers: Please be familiar with the time and location of your talk and stream. Arrive at the location of your talk 10 mins prior to the scheduled start time. Introduce yourself to the chair of the session / stream organiser. With the help of the chair, familiarise yourself with the equipment in the room and how to connect your laptop. Presenters should bring a laptop with them as it is very expensive to hire one. 14 Talks are strictly 20 mins long with 5 mins for questions & answers (except for Keynote/Tutorial presentations) The stream chair will give indications of time remaining. You are encouraged to also bring your presentation on a USB stick or CD/DVD as a backup. For Stream Chairs Arrive at the appropriate seminar room 10 mins before the start of the stream session. Familiarise yourself with the equipment in the room, including how to connect laptops to the projector. Introduce yourself to your speakers. Your main role will be to ensure that the stream runs to time. Each talk (except keynotes and tutorials) is scheduled for 20 mins plus 5 mins for Q&A. There is a 5 mins break to allow movement of delegates between stream rooms. Equipment Equipment in each presentation room will include Data Projector; Projector Screen; DVD player if required (please advise if this is needed!) Messages/Announcements All timetable announcements as well as messages for delegates will be displayed on a notice board in the foyer of the JMCC. Please make sure that you check regularly for any new information. Social Events Tuesday evening will see our traditional Bar Quiz in the JMCC Prestonfield room (1st Floor) from 21:00 ” Midnight, ( with the help of our esteemed compere Amanda Tucker of Dstl!). A cash bar will be open in the Centro Bar, near the Prestonfield room from 19:30 ” Midnight. Wednesday afternoon social trips to: Glenkinchie Whisky Distillery Lauriston Castle National Museum of Flight (Please note that there are a lot of stairs at Lauriston Castle and no wheelchair access). Coaches will start loading at 14:45 for a sharp departure at 15:00. Coaches will return to this area between 17:30 ” 18:30. Coaches leave from (and return to), outside the entrance to the Edinburgh First building on Holyrood Road at 15:00 sharp! Wednesday evening drinks receptions: We are in the happy position of having two drinks receptions on Wednesday (with time for a coffee in between for those lucky people who attend both!). Making an Impact stream: Delegates attending the Making an Impact Practitioner sessions on Wednesday are invited to enjoy a glass of wine in the JMCC Centro bar area (1st Floor), at the close of day at 18:00. (Kindly sponsored by SIMUL8 Corporation.). The OR Society ” pre-dinner drinks: A drinks reception for all delegates will be held in the St. Trinneans room in St. Leonards Hall from 19:15 20:00. 15 Palgrave Macmillan, the main sponsor of this year’s drinks reception will be saying a few words about their celebration of the 10th anniversary of KMRP, the 25th of OR Insight, and the launch of Health Systems. We are also very pleased to welcome a new sponsor, 2020 Consulting, who have also contributed to this event. Wednesday evening Gala Dinner and dance: Will follow the OR Society drinks reception and be served in South Hall (a few minutes’ walk from St. Leonards Hall) followed by Ceilidh dancing to a live band. Please bring your Gala Dinner ticket with you as it will be collected at the door prior to entry. 20:00 ” 22:00 A gala dinner will be served in the Banqueting Suite in South Hall. 22:00 ” Midnight. The banquet will be followed by Ceilidh dancing (or listening!) to Matt Smith and The Little Biggar Band. A cash bar will be open in South Hall 20:00 - Midnight. A quieter area is available upstairs in South Hall for those that wish to talk (still near to the bar!). Alternatively the JMCC bar on the ground floor of the JMCC building will be open until 11:00 (although this is not exclusively for OR Society use). Beers available are Draught ” Carling Lager/ Belhaven Best. Bottled ” Miller, Peroni, and Magners etc. Medical Assistance If in the conference venue, contact the Duty Manager. Otherwise dial 6666 from your bedroom to speak to switchboard or security staff. (Details are also noted in an information booklet within your bedroom.) The telephone number for colleagues or family to leave an urgent message for delegates is: Tel: +44 (0)131 667 1971 Facilities: Internet – The JMC Centre, St. Leonards Hall and the Edinburgh First reception area, have free wireless Internet access. All delegates will be provided with a wifi pass. Please collect from the OR Society reception desk when you register. This can be used to access the internet in all the public areas and meeting rooms. There is also wireless access in the quiet rooms on the corridors of Chancellors Court. Wired Internet access (not wifi), is available in the bedrooms. The charges are as follows:1 Day = £5, 2 ” 3 Days = £10 or 4 ” 7 Days = £15.00. Contact Edinburgh First reception for further information. (A data cable is in each of the bedrooms for guests to use, although you may wish to bring your own). Please bring your own laptops with you, (as they are very expensive to hire) and presentations on a USB or similar as back-up. 16 Taxi Numbers ” A direct free phone is available within the Edinburgh First Reception building to Central Taxis. Shop - A shop is available on the ground floor of the JMCC building which sells newspapers refreshments and food etc. Open daily 0800 ” 1800 hrs. Chemist ” The nearest Chemist is Boots, 28 ” 30 Newington Road. Tel: 0131 667 3008. Laundry facilities - are available within the Chancellor’s Court Building. Banks - There is an ATM on the outside of the JMCC building and information re local banks will be noted in the information booklet in your bedroom. Bank of Scotland, 51, Clerk Street. Tel: 0131 465 3414. Clydesdale Bank, 29-31 Nicolson Square. Tel: 0845 724 0024. Sports Facilities - The Pollock Halls are situated immediately behind the Royal Commonwealth Swimming Pool. Guests can receive special rates for use of all the facilities (including swimming pool & multi-gym. If you have a bedroom key you will need to show this. All details are available from Edinburgh First Reception. 17 OR54 Has Gone Mobile! The OR Society’s Annual Conference, OR54, has gone mobile using Guidebook! We strongly encourage you to download our mobile guide to enhance your experience at OR54. You'll be able to plan your day with a personalised schedule and browse exhibitors, maps and general conference info. The app is compatible with iPhones, iPads, iPod Touches and Android devices. Windows Phone 7 and Blackberry users can access the same information via our mobile site at m.guidebook.com. To get the guide, choose one of the methods below: 1. Download 'Guidebook' from the Apple App Store or the Android Marketplace 2. Visit http://guidebook.com/getit from your phone's browser 3. Scan the following image with your mobile phone (QR-Code reader required, e.g. 'Red Laser', 'Barcode Scanner') The guide will be listed under the "download guides" section of the application. Search for ‚OR54‛ or select Redeem Code and type in xxxxxx 18 Tuesday 04 September 2012 Time JM Pentland 08:00 08:30 09:00 09:30 10:00 10:30 JM Prestonfield JM Salisbury JM Holyrood JM Duddingston ST.L. Pollock ST.L. Nelson Community OR (OR54A744) Health (OR54A531) Problem Structuring Methods (OR54A541) Simulation (OR54A562) Stochastic Scheduling & DAP (OR54A619) Community OR (OR54A605) Health (OR54A646) Problem Structuring Methods (OR54A621) Simulation (OR54A699) Stochastic Scheduling & DAP (OR54A633) Health (OR54A688) Problem Structuring Methods (OR54A752) Simulation (OR54A732) Stochastic Scheduling & DAP (OR54A669) 12:30 13:00 ST.L. Cowan ST.L. Brewster JM Pentland East JM Pentland West OR in Education (OR54A575) OR in Education (OR54A721) Forecasting, Data Mining & CIM (OR54A585) OR & Strategy (OR54A524) Info.Systems & Knowledge Mgmt (OR54A553) Routing Applic's & Transport (OR54A529) Forecasting, Data Mining & CIM (OR54A612) OR & Strategy (OR54A675) Project Management (OR54A528) Routing Applic's & Transport (OR54A599) Forecasting, Data Mining & CIM (OR54A683) OR & Strategy (OR54A733) Project Management (OR54A668) Routing Applic's & Transport (OR54A638) Project Management (OR54A618) Metaheuristics (OR54A548) Project Management (OR54A749) Metaheuristics (OR54A625) Project Management (OR54A707) Metaheuristics (OR54A680) OperationsDesignInnovation Interface (OR54A654) OperationsDesignInnovation Interface (OR54A657) Lunch 12:30 - 13:30 in JM Health (OR54A656) Problem Structuring Methods (OR54A579) Optimisation (OR54A588) Scheduling (OR54A559) Analytics (OR54A751) Forecasting, Data Mining & CIM (OR54A573) Health (OR54A705) Problem Structuring Methods (OR54A620) Optimisation (OR54A685) Scheduling (OR54A660) Analytics (OR54A757) Forecasting, Data Mining & CIM (OR54A570) Optimisation (OR54A714) Scheduling (OR54A591) 13:30 OR & Strategy (OR54A644) OR & Strategy (OR54A659) 14:30 15:00 ST.L. Bonnar Welcome Address: 09:30am (JM - Pentland) - Tom Archibald, Chair. Opening Plenary: Dr Geoff Royston, President of The OR Society - 'Operational Research for the real world: big questions from a small island' 11:30 14:00 ST.L. St.Trinneans Registration in JM 11:00 12:00 Conference Programme at a Glance Refreshments in JM 19 Info.Systems & Knowledge Mgmt (OR54A561) Info.Systems & Knowledge Mgmt (OR54A568) Info.Systems & Knowledge Mgmt (OR54A712) Horizon Community OR Scanning & (OR54A616) Future Analysis (OR54A696) Horizon Community OR Scanning & (OR54A617) Future Analysis (OR54A745) Horizon Community OR Scanning & (OR54A649) Future Analysis (OR54A748) Tuesday 4 September 2012 continued Time 15:30 16:00 JM Pentland Community OR (OR54A693) JM Prestonfield Health (OR54A652) Community OR (OR54A739) Health (OR54A730) Health (OR54A655) 16:30 JM Salisbury Problem Structuring Methods (OR54A518) Problem Structuring Methods (OR54A725) Problem Structuring Methods (OR54A736) Conference Programme at a Glance JM Holyrood Optimisation (OR54A571) JM Duddingston Scheduling (OR54A533) ST.L. St.Trinneans Logistics and Supply Chain (OR54A560) ST.L. Pollock Analytics (OR54A667) ST.L. Nelson OR & Strategy (OR54A521) ST.L. Bonnar Project Management (OR54A715) ST.L. ST.L. Cowan Brewster Routing Applic's Green Logistics & Transport (OR54A532) (OR54A572) Optimisation (OR54A596) Scheduling (OR54A567) Logistics and Supply Chain (OR54A679) Analytics (OR54A756) OR & Strategy (OR54A665) Routing Applic's Green Logistics & Transport (OR54A663) (OR54A628) Optimisation (OR54A664) Scheduling (OR54A661) Logistics and Supply Chain (OR54A690) Analytics (OR54A743) OR & Strategy (OR54A658) Routing Applic's & Transport (OR54A632) Plenary: 17:00 (JM Pentland) John Hopes, Ernst & Young - 'OR and analytics – an opportunity for growth' 18:00 Government Session - (JM Pentland) - Vivienne Raven, GORS - 'Recruitment to the Government Operational Research Service' 18:30 Open Meeting to discuss OR Society matters - (JM Pentland) – Geoff Royston and Gavin Blackett Time JM Pentland 09:00 JM Pentland East JM Pentland West Conference Programme at a Glance JM Prestonfield JM Salisbury JM Holyrood JM Duddingston Health (OR54A606) OR in Education (OR54A629) Green Logistics (OR54A545) OR in Education (OR54A703) Green Logistics (OR54A547) Making an Impact 08:20 ” 08:30 Introduction and Welcome 08:30 ” 09:30 Speed Networking 08:00 08:30 JM Pentland West Stochastic Scheduling & DAP (OR54A624) 17:00 Wednesday 05 September 2012 JM Pentland East ST.L. ST.L. ST.L. St.Trinneans Pollock Nelson Registration in JM Logistics and Forecasting, OR & Strategy Supply Chain Data Mining (OR54A639) (OR54A534) & CIM (OR54A750) OR & Strategy (OR54A595) ST.L. Bonnar Optimisation (OR54A641) 09:30 10:00 Plenary: 09:30 (JM Pentland) Jason Field, Dstl - 'Operational Research – Supporting Defence and Security' 10:30 Refreshments in JM 20 ST.L. Cowan ST.L. Brewster Metaheuristics (OR54A697) Queue Modelling (OR54A631) Metaheuristics (OR54A710) Queue Modelling (OR54A706) Wednesday 05 September 2012 continued Time JM Pentland JM Prestonfield Health (OR54A610) 11:00 Health (OR54A666) 11:30 JM Salisbury Making an Impact Technique Tasters 2 sessions of 45 minutes each 11:00-11:45 (OR54A763) 11:45-12:30 (OR54A764) Health (OR54A695) 12:00 Conference Programme at a Glance JM Holyrood Simulation (OR54A564) JM Duddingston Optimisation (OR54A577) ST.L. St.Trinneans Project Management (OR54A650) ST.L. Pollock Forecasting, Data Mining & CIM (OR54A587) ST.L. Nelson OR & Strategy (OR54A535) Simulation (OR54A582) Optimisation (OR54A603) Logistics and Supply Chain (OR54A602) OR & Strategy (OR54A598) Simulation (OR54A592) Optimisation (OR54A729) Logistics and Supply Chain (OR54A691) Forecasting, Data Mining & CIM (OR54A640) Forecasting, Data Mining & CIM (OR54A740) OR & Strategy (OR54A586) ST.L. Bonnar Making an Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A767) Making an ST.L. Cowan Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A768) Making an ST.L. Brewster Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A769) Making an Making an Impact ” Additional Workshops that may run depending on demand: 1.Horizon Scanning; sharing experiences 2.Ethical Dilemmas in OR 3.Selling OR 12:30 13:15 Lunch 12:30 - 13:15 = 45 minutes 13:15 13:45 14:15 President's Medal presentations 13:15 - 14:45 (JM Pentland) 13:15 - 'From Operational Research to Operational Planning: Shaping the NATO plan for Afghanistan', Patrick Rose, Colin Marston, Dstl 13:45 - 'Modelling demand and performance in HMRC call centres', Stephen O'Donnell, HMRC 14:15 - 'Decision Analysis in Nuclear Decommissioning: Developments in Techniques and Stakeholder Engagement Processes 14:45 15:30 Making an Impact Technique Tasters (OR54A774) Making an Impact Technique Tasters (OR54A773) Making an Impact Technique Tasters (OR54A772) Making an Impact Technique Tasters (OR54A771) Making an Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A775) Making an Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A776) Making an Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A777) JM Pentland East Impact ” Facilitated Workshops on 'Hot' Topics in OR (OR54A770) Making an Impact ” Additional Workshops that may run depending on demand: 1.Horizon Scanning; sharing experiences 2.Ethical Dilemmas in OR 3.Selling OR 4.Personal and Professional Development 16:15 Refreshments in JM 16:30 Making an Impact: Practitioner – Academic Collaboration: Maximising Research Impact 16:30 - 17:45 (JM Pentland) - Chaired by Sally Brailsford, University of Southampton. Practitioners - Sean Jones (NATS), Tony O'Connor (GORS), Stewart Robinson (Loughborough) and Mike Pidd (Lancaster). 17:45 Review of the Day followed by a Drinks Reception for MAI delegates at 18:00 sponsored by SIMUL8 Corporation 19:15 19:15 Drinks Reception in the St. Trinnean Room in St. Leonards Hall sponsored by Palgrave Macmillan with part-sponsorship from 2020 Consulting. Our Gala Dinner will be held in South Hall from 20:00 21 JM Pentland West Making an Impact Technique Tasters 2 sessions of 45 minutes each 11:00-11:45 (OR54A765) 11:45-12:30 (OR54A766) Thursday 06 September 2012 Time JM Pentland Conference Programme at a Glance JM Prestonfield JM Salisbury JM Holyrood JM Duddingston Health (OR54A651) Credit Risk (OR54A557) Simulation (OR54A556) Green Logistics (OR54A542) Health (OR54A662) Credit Risk (OR54A692) Green Logistics (OR54A549) Health (OR54A673) Credit Risk (OR54A734) Green Logistics (OR54A614) 08:00 08:30 09:00 09:30 10:00 ST.L. St.Trinneans ST.L. ST.L. Pollock Nelson Registration in JM Forecasting, OR Consultancy Data Mining /Case Studies & CIM (OR54A678) (OR54A550) Logistics and Forecasting, OR Consultancy Supply Chain Data Mining /Case Studies (OR54A611) & CIM (OR54A682) (OR54A565) Forecasting, OR Consultancy Data Mining /Case Studies & CIM (OR54A687) (OR54A584) Refreshments in JM ST.L. Bonnar ST.L. Cowan ST.L. Brewster Queue Modelling (OR54A594) Operations/Design /Innovation Interface (OR54A674) Operations/Design /Innovation Interface (OR54A681) Stochastic Inventory Control (OR54A544) Stochastic Inventory Control (OR54A623) Queue Modelling (OR54A642) JM Pentland East JM Pentland West Queue Modelling (OR54A698) Plenary: 10:30 (JM Pentland) - Chris Dent, Durham University 'Challenges in energy system modelling: data, optimisation and futurology' 10:30 11:00 11:30 12:00 Health (OR54A726) Credit Risk (OR54A636) Health (OR54A702) Credit Risk (OR54A637) Simulation (OR54A630) Green Logistics (OR54A539) Green Logistics (OR54A709) 12:30 13:00 13:30 14:00 14:30 15:00 Stochastic Inventory Control (OR54A554) Stochastic Inventory Control (OR54A626) Forecasting, Data Mining & CIM (OR54A540) Forecasting, Data Mining & CIM (OR54A700) OR Consultancy /Case Studies (OR54A601) Optimisation (OR54A597) Operations/Design /Innovation Interface (OR54A672) OR Consultancy Routing Applic's Operations/Design /Case Studies & Transport /Innovation (OR54A731) (OR54A647) Interface (OR54A741) Lunch 12:30 - 13:30 in JM Health (OR54A566) Credit Risk (OR54A546) Simulation (OR54A537) Green Logistics (OR54A615) Health (OR54A593) Credit Risk (OR54A558) Simulation (OR54A676) Green Logistics (OR54A738) Credit Risk (OR54A627) Simulation (OR54A713) Logistics and Supply Chain (OR54A607) Forecasting, OR Consultancy Routing Applic's Operations/Design Data Mining & /Case Studies & Transport /Innovation CIM Presentations (OR54A635) Interface (OR54A583) x4 of 22.5 (OR54A670) minutes each Routing Applic's Operations/Design Logistics and Forecasting, 13:30 Supply Chain Data Mining & & Transport /Innovation (OR54A580) (OR54A608) CIM (OR54A689) Interface 13:52 (OR54A590) (OR54A677) (OR54A701) Logistics and Forecasting, Routing Applic's Operations/Design 14:14 Supply Chain Data Mining & & Transport /Innovation (OR54A737) (OR54A634) CIM (OR54A735) Interface 14:36 (OR54A686) (OR54A653) (OR54A684) Conference closes at 15:00 22 Optimisation (OR54A578) Optimisation (OR54A589) Tuesday 04 September 2012 Time JM Pentland 08:00 08:30 09:00 09:30 10:00 10:30 11:00 JM Prestonfield FULL CONFERENCE PROGRAMME JM Salisbury JM Holyrood JM Duddingston ST.L. St.Trinneans ST.L. Pollock ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) ST.L. Bonnar ST.L. Cowan ST.L. Brewster JM Pentland East Registration in JMCC Welcome Address: 09:30am (JM - Pentland) - Tom Archibald, Chair. Opening Plenary from Dr Geoff Royston, President of The OR Society - 'Operational Research for the real world: big questions from a small island' Community OR Health Problem Simulation Stochastic Structuring Scheduling & KEYNOTE: Operational Forward Methods DAP Modelling To Research Monte-Carlo Rich Pictures For American The Support At The Evaluation Of Communities Welsh Govern- Tessa Berg Options Risk In Gerard de Daniel Weiment Zeeuw and Chung Miao and PharmacTracey England, Martha Vahl Paul Harper and Yung-Hsin Lee eutical Sarah Lowe Research Anne-Marie Oreskovich and John Gittins Community OR Cop Approach To Evaluating Expert Support 11:30 Ricardo Barros, Gerald Midgley and Luis Pinzón Health Balancing Beds And Theatres John Bowers Problem Structuring Methods Schools Of Thought Analysis (SOTA) Fred Cameron and Geoff Pond Simulation Empirical Bayes Methods For DES Stochastic Scheduling & DAP Inventory Optimization Via ADP Shona Blair, Tim Bedford and John Quigley Chris Kirkbride OR in Education OR In Schools Forecasting, Data Mining & CIM Vincent Knight and Louise Orpin KEYNOTE: Forecasting Most Successful Products OR & Strategy Generic Models For Start-Ups Martin Kunc Fotios Petropoulos, Vassilios Assimakopoulos, Akrivi Litsa, Konstantinos Nikolopoulos and Yiannis Polychronakis Forecasting, OR & Strategy Data Mining Competitive & CIM Archetypes Demand Yan Emma Liu Forecast and Shu-Jung Considering Media Effects Sunny Yang Ryunosuke Usami and Ichie Watanabe 23 Info.Systems &Knowledge Mgmt Event Monitoring A Large Financial Enterprise Caroline Sanders Routing Applic's & Transport Metaheuristics For Vehicle Routing Problems Prasanna Kumar Ramaswamy and Jamal Ouenniche Project Routing Applic's Management & Transport Cultural Dynamic Diversity And Congestion Workplace Charging Kathryn Stewart Dynamics James Freeman Operations DesignInnovation Interface Integrated Healthcare Network Facilitators Jiun-Yu Yu Operations DesignInnovation Interface Modelling Alternate Bargaining Andrew Brint, Andrea Genovese and Niraj Kumar JM Pentland West Tuesday 04 September 2012 continued Time JM Pentland JM Prestonfield JM Salisbury FULL CONFERENCE PROGRAMME JM Holyrood JM Duddingston ST.L. St.Trinneans ST.L. Pollock ST.L. Nelson Health Designing Quota Size And Patient Arrival Time 12:00 Problem Simulation Stochastic OR in Education Forecasting, OR & Strategy Structuring Scheduling Simulation In OR Outreach Data Mining & Supporting Methods & DAP CIM The Supply At British InternationalNATO Dynamic Long-Term isation Using Chain Domain Airways SAS-087 Pricing In Energy Chris Owen, Naomi Crouch SD Presented By Pavel Albores Yield And Forecasting:Th Juan Pablo Kwei-Long and Doug Love Huang and JeiNeville J Revenue e Case Of EU Torres Fen Tsai Nikoletta Koleri, Curtis Management Neville Curtis ST.L. Cowan Project Management Intellectual Framework Routing Applic's & Transport Vehicle Routing With Dependent Vehicles Gary Bell and Jon Warwick ST.L. Brewster JM Pentland East JM Pentland West Edward Kent and Jason Atkin Lunch 12:30 - 13:30 in JMCC Health Information Design In Health Communications 13:30 ST.L. Bonnar Dimitrios Askounis and Charikleia Karakosta Michael Pearson 12:30 13:00 (please note that authors names are listed in the order of Presenter first and then alphabetically) Problem Optimisation Scheduling Structuring Solving Bicriteria Methods Constraint Forge Teaching Programming Scheduling Problem Problems With Rasaratnam Structuring Logendran and AIMMS Yasaman Methods Guido Diepen Sabrina Pei-Yi Cheng and Fran Ackermann and James Jiun-Yu Yu Alexander Mehravaran Analytics R For Analytics Forecasting, OR & Strategy Data Mining Modelling & CIM Degradation Chris Campbell Forecasting Of Control & Policy Monitoring ImplementAssets ation Graeme Blair, Strategies John Quigley and Konstantinos Nikolopoulos, Vassilis Asimakopoulos, Akrivi Litsa and Fotios Petropoulos 24 Lesley Walls Project Management Project Success: An Ethical Perspective Rosane Pagano and I. K. White Metaheuristics Variable Neighbourhood Structures John Lamb Info.Systems Horizon Community OR &Knowledge Scanning & OR In The 3Rd Mgmt Future Analysis Sector Weaving The The Ten Ruth Kaufman Innovation CommandWeb ments Of Jo Smedley Horizon Scanning John Carney Tuesday 04 September 2012 continued Time 14:00 JM Pentland JM Prestonfield JM Salisbury Health General Practitioner Funding Formula Problem Structuring Methods Strategic Choice To Orville D'Silva, Analyse Naval David Deployment Worthington and Edward Terry James Crosbie Seagriff FULL CONFERENCE PROGRAMME JM Holyrood JM Duddingston ST.L. St.Trinneans ST.L. Pollock Optimisation Scheduling Analytics Forecasting, Stochastic Scheduling Big Brother Is Data Mining & CIM Optimisation RAF Air Traffic Still Watching, Wind Power Of Reliability Control But ... Quantile Jacqui Taylor Development Personnel Forecasting Kerem Richard Conniss, Jooyoung Jeon and James Taylor Akartunali and Tim Curtois and John Quigley Sanja Petrovic Optimisation GD Based For HHMO Scheduling A Genetic Algorithm ST.L. Bonnar ST.L. Cowan JM JM Pentland Pentland East West Project Metaheuristics Info.Systems Horizon Community OR Management &Knowledge Scanning & Heuristic Mitigating Mgmt Future Analysis Methodology Function Potential Preparing Conceptual For Simulation Resynthesis Conflicts In Force Conceptual Jerry Swan, Your Data For Sustainability Edmund Burke, Advanced Development Eliseo VilaltaModelling Hannah Locke Miles Weaver, Graham Kendall Analytics Perdomo Pavel Albores and Ender Ozcan and Doug Love OR & Strategy Data Analysis In The Water Industry Mashael Maashi, Khaled Alhamad Graham Kendall and Ender Ozcan 14:30 ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) Efthalia Anagnostou and Deborah Gee Project Management Holon Rich Pictures Gary Bell and Jon Warwick Metaheuristics InstanceSpecific Parameter Tuning In Practice Jana Ries and Patrick Beullens ST.L. Brewster Sayara Beg Info.Systems Horizon &Knowledge Scanning & Mgmt Future Analysis KM And Systems, Sustainable Scanning, And Quality Scenarios Brian Lehaney Wendy L Schultz Refreshments in JMCC 15:00 Community OR Health Problem Structuring Researching Melanoma Methods ‘Wicked’ Pathways OR For Air Issues Flight Military ‘Messy’ World Simulator Assistance To Paul Holmstrom, Of COR 15:30 Magdalena Security And Jennifer Claesson and Jackson Dev Stefan Hallberg John Holt Optimisation Scheduling Logistics and Analytics OR & Strategy Supply Chain Models For Air Competitive Royal Mail Is There Still Acquisition & Cargo Travelling Goes To Life In Acceptance The Olympics! Revenue Salesmen Stafford Risks Paul Liddiard Beer's VSM? Management Problem In Military Emily Cookson, Jiawei Li Stephen Kevin Harwood Systems Glazebrook and Joern Meissner Patrick Driscoll, Timothy Elkins and James Enos 25 Project Routing Applic's Green Logistics Management & Transport The VRP With KEYNOTE: Ant Colony Multiple Lessons For Optimization Driving Project For A School Ranges Management Bus Routing Angel A. Juan, Colin Eden and Jose Jimenez, Tolga Bektas, Susan Howick Jairo R Montoya- Javier Faulin Torres and Juan and Jarrod Sebastian AriasGoentzel Rojas Community OR PSM & Project Management In Community Regeneration Michael Charlton Tuesday 04 September 2012 continued Time JM Pentland JM Prestonfield JM Salisbury Community OR Engineering Tests Of TASER Devices Health Mechanism Design For Examination Resource Alloc. Problem Structuring Methods A Decision Making Frame-work 16:00 Fred Cameron, FULL CONFERENCE PROGRAMME JM Holyrood Optimisation On A WarmStarted Primal-Dual Column Generation I-Hsuan Hong, Julian Benavides Me Jack Brimberg, Bill Hurley, Bill Wen-Chih Chen Franco and Juan Pablo GonzalezSimms and and Kun-Liang Felipe Henao Brevis and Jacek Xiaofeng Yang Lien Gondzio 16:30 17:30 18:00 18:30 ST.L. St.Trinneans Scheduling Logistics and A Look Ahead Supply Chain Heuristic For Supply Chain Competition MOSP Marco Carvalho And Capacity and Nei Soma Constra-ints ST.L. Pollock ST.L. Nelson Analytics Enter The Matrix OR & Strategy Integrating Behaviour Change Into OR Ian Taylor Philip Jones Shu-Jung Sunny Yang, Eddie Anderson and Yong Bao Akartunali and Vicky Mak-Hau Andreas Grothey and Ken McKinnon and Gilbert Laporte ST.L. Bonnar ST.L. Cowan OR & Strategy Social Media And Engagement With Scenario Projects Paul Lam, Paul Frances O'Brien Ingles and and Maureen Thibaut Sacreste Meadows ST.L. Brewster Routing Applic's Green Logistics & Transport Grocery Variants Of Deliveries The From Depot Capacitated To Arc Routing Store Problem Daniel Black Gu Pang and Luc Muyldermans Health Problem Optimisation Scheduling Logistics and Analytics Structuring Supply Chain Improving Radiation Stochastic Agile Methods Performance Treatment Unit Commit- The Orienteer- Approach To ing Problem Of Emergency Issue IdentificPlanning ment By Knowledge ation At The Optimiz-ation With Variable Medical Column Discovery Front-End Of Profits Service For VMAT Generation Of Web A Study Jiun-Yu Yu Kerem Tim Schulze, Gunes Erdogan Log Neville Curtis 17:00 JM Duddingston (please note that authors names are listed in the order of Presenter first and then alphabetically) and Richard Eglese Routing Applic's Stochastic & Transport Scheduling & DAP Vehicle Index Policies Routing Problem With For Stochastic Machine Driver Familiarity Maintenance Matthew Soulby Diego Ruizand Jason Atkin Hernandez and David DelgadoGómez Plenary: 17:00 (JM Pentland) John Hopes, Ernst & Young 'OR and analytics – an opportunity for growth' Government Session - (JM Pentland) - Vivienne Raven, GORS - 'Recruitment to the Government Operational Research Service' Open Meeting to discuss OR Society matters - (JM Pentland) Geoff Royston and Gavin Blackett 26 JM Pentland East JM Pentland West Wednesday 05 September 2012 Time JM Pentland JM Prestonfield 08:00 Health KEYNOTE: OR And The Industrialisation Of Healthcare 08:30 Michael Pidd FULL CONFERENCE PROGRAMME JM Salisbury JM Holyrood JM Duddingston ST.L. St.Trinneans ST.L. Pollock Registration in JMCC OR in Education Green Logistics Making an Logistics and Forecasting, OR & Strategy Impact Supply Chain Data Mining & Statistical The TimeGroup CIM RemanufactModel To Dependent Dynamics And 08:20 ” 08:30 Predictive Mental Model uring With Estimate Pollution Introduction Maintenance Convergence Multiple Quality Routing And Welcome Collection Modelling Kenneth Research Problem With Dbns Kyunghyun Sites Alberto PaucarAnna Ken McNaught Huh Caceres and Xin Franceschetti, 08:30 ” 09:30 Christos Speed and Adam Shi Gilbert Laporte, Zikopoulos Zagorecki Tom van Networking and George Woensel, Tolga Bektas and Dorothee Honhon (Facilitated by Sophie Carr) ST.L. Bonnar ST.L. Cowan ST.L. Brewster Metaheuristics Queue Modelling HyperCo-Ordinated Heuristics Queues At For Magic Arterial Squares Intersections Ahmed Kheiri César Dario and Ender Ozcan Velandia, Jason Atkin, Ruibin Bai and Graham Kendall Tagaras OR in Education Green Logistics Modelling Fuel Latent Emissions Preferences Optimization Martin Spollen, In Vrps 09:00 ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) OR & Strategy Optimisation Metaheuristics Queue Modelling Strategic GRS Model In A Hybrid GA When To Use Management, DEA For The TRP Virtual Hold Cagla Role Playing Miki Tsutsui and Technology Kaoru Tone Cergibozan and David &System A. Serdar Tasan Worthington and Dynamics Karen Cairns and Richard Eglese Adele Marshall and Jiani Qian Federico Barnabè, Cristiano Busco, Pål I. Davidsen, Maurizio Lambri and Gianfranco Zatta 09:30 10:00 Plenary: 09:30 (JM Pentland)- Jason Field, Dstl - 'Operational Research – Supporting Defence and Security' 10:30 Refreshments in JMCC 27 Chris Kirkbride JM Pentland East JM Pentland West Wednesday 5 September 2012 continued Time 11:00 JM Pentland JM Prestonfield JM Salisbury Health Monitoring Outcomes In Paediatric Heart Surgery Making an Impact Technique Tasters 2 sessions Christina Pagel, of 45 Kate Brown, minutes each Sonya Crowe and Martin Utley Health Sequential Drug Decision Problems In Healthcare 11:30 12:00 Eun-Ju Kim, Roberta Ara, Jim Chilcott and Yang Meng Health Soapt: Service Option Assessor And Prioritisation Carol Marshall 11:00 -11:45 Rich Pictures Giles Hindle 11:45 -12:30 Systems Modelling Giles Hindle FULL CONFERENCE PROGRAMME JM Holyrood JM Duddingston ST.L. St.Trinneans ST.L. Pollock Simulation Optimisation Project Forecasting, Data Mining Current Deployment Management & CIM Planning With Practice Of Of ICT ExperimenServices In The OI Factor A Comparison tation In D.E.S Cloud Data Michelle Morris Of Relative Accuracy Kathryn Hoad, Centres Tom Monks and Bjorn Nygreen Measures Frances O'Brien Christopher Tofallis and Anders Nordby Gullhav Simulation Optimisation Modeling Matrix-Free Assembly Line Interior Point Using Method Kimon Simulation Logistics and Supply Chain Developing Simplified And Linearised Models Forecasting, Data Mining & CIM Mining Pattern Behavior Norhanom Awang and Razman Mat Tahar Fountoulakis, Jacek Gondzio and Pavel Zhlobich Simulation Modeling And Simulation Optimisation Logistics and Forecasting, Supply Chain Data Mining Reilability & CIM A Generalised And Anomaly Maintenance Multi-Echelon Detection In Inventory Of Complex Point Clouds Theory Plant Fazeeda Mohamad and Razman Mat Tahar Virginia Luisa Marques Spiegler, Mohamed M. Naim, Denis R. Towill and Joakim Wikner Norhafiza Ahmad Ahmed AlHinai Patrick Beullens, Dafydd Evans Gerrit K. and Babakalli Janssens and Luk Alkali N. Van Wassenhove ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) ST.L. Bonnar ST.L. Cowan OR & Strategy Making an Making an Impact Impact Metaphorical Facilitated Facilitated Models And Workshops Workshops Credible on 'Hot' Topics on 'Hot' Topics Worlds in OR in OR John Morecroft Networking Build Your For Introverts Own Pub! Frances Sneddon Dave Buxton ST.L. Brewster JM Pentland East Making an Making an Impact Impact Facilitated Facilitated Workshops Workshops on 'Hot' Topics on 'Hot' Topics in OR in OR Data Personal And Visualisation Professional Jacqui Taylor Development K. Swatridge OR & Strategy Systems Thinking Applied To Triple Helix Theories UngKyu Han and Martin Kunc Making an Impact ” Additional Workshops that may run depending on demand: 1.Horizon Scanning; sharing experiences 2.Ethical Dilemmas in OR 3.Selling OR OR & Strategy Assessing Strategic Risk In The Public Sector Ian Birchmore 12:30 13:15 Lunch 12:30 - 13:15 = 45 minutes 13:15 13:45 14:15 President's Medal presentations 13:15 - 14:45 (JM Pentland) 13:15 - 'From Operational Research to Operational Planning: Shaping the NATO plan for Afghanistan', Patrick Rose, Colin Marston, Dstl 13:45 - 'Modelling demand and performance in HMRC call centres', Stephen O'Donnell, HMRC 14:15 - 'Decision Analysis in Nuclear Decommissioning: Developments in Techniques and Stakeholder Engagement Processe 28 JM Pentland West Making an Impact Technique Tasters 2 sessions of 45 minutes each 11:00 -11:45 Bayesian Methods Sophie Carr 11:45 -12:30 Constraint Programming Guido Diepen Wednesday 5 September 2012 continued Time JM Pentland Making an Impact Technique Tasters 14:45 Rich Pictures Giles Hindle JM Prestonfield Making an Impact Technique Tasters Bayesian Methods Sophie Carr 15:30 Making an Impact Technique Tasters Systems Modelling Giles Hindle Making an Impact Technique Tasters Constraint Programming JM Salisbury FULL CONFERENCE PROGRAMME JM Holyrood JM Duddingston Making an Making an Making an Impact Impact Impact Facilitated Facilitated Facilitated Workshops on Workshops on Workshops on 'Hot' Topics in 'Hot' Topics in 'Hot' Topics in OR OR OR Build Your Data Networking Own Pub! Pm Visualisation For Introverts David Buxton Pm Pm ST.L. St.Trinneans ST.L. Pollock ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) ST.L. Bonnar ST.L. Cowan ST.L. Brewster JM Pentland East Making an Impact ” Additional Workshops that may run depending on demand: 1.Horizon Scanning; sharing experiences 2.Ethical Dilemmas in OR 3.Selling OR 4.Personal and Professional Development Jacqui Taylor Frances Sneddon Guido Diepen 16:15 Refreshments in JM 16:30 Making an Impact: Practitioner – Academic Collaboration: Maximising Research Impact 16:30 - 17:45 (JM Pentland) - Chaired by Sally Brailsford, University of Southampton. Practitioners - Sean Jones (NATS), Tony O'Connor (GORS), Stewart Robinson (Loughborough) and Mike Pidd (Lancaster). 17:45 Review of the Day followed by a Drinks Reception for MAI delegates at 18:00 sponsored by SIMUL8 Corporation 19:15 19:15 Drinks Reception in the St. Trinnean Room in St. Leonards Hall sponsored by Palgrave Macmillan with part-sponsorship from 2020 Consulting. Our Gala Dinner will be held in South Hall from 20:00 29 JM Pentland West Thursday 06 September 2012 Time JM Pentland JM Prestonfield FULL CONFERENCE PROGRAMME JM Salisbury JM Holyrood JM Duddingston 08:00 Health Credit Risk Simulation Green Logistics Simulation For Survival TUTORIAL: On Dynamic Sustainability Analysis Conceptual Collection In Healthcare Jonathan Crook, Modelling Scheduling Masoud Fakhimi, Tony Bellotti and Fraser McLeod, For Mindy Leow Navonil Tolga Bektas, Simulation 08:30 Mustafee and Jane Probert Stewart Robinson Tom Cherrett and Gunes Erdogan ST.L. St.Trinneans ST.L. Pollock ST.L. Nelson (please note that authors names are listed in the order of Presenter first and then alphabetically) ST.L. Bonnar Registration in JMCC Forecasting, OR Consultancy Queue Modelling OperationsData Mining /Case Studies DesignPrice Of & CIM Innovation Decision Anarchy In Interface Modelling & Analysis Queueing Infection Forecasting Model Of Systems Control With PharmacDisease Rob Shone Strategic eutical Life Management Healthcare Cycles John Soje Worker Samantha 09:00 Paul Holmstrom, Marie Elf, Inga Malmqvist, Kerstin Öhrn and Lena von Koch Health OverAnalysis In Health Care And Health Services 09:30 Penelope Mullen Credit Risk Joint Credit Losses And Poisson Processes Fernando Moreira Credit Risk Models Of Financial Distress In Chinese Firms Anthony Bellotti and Xijuan Bellotti ST.L. Brewster Stochastic Inventory Control SDP In Perishable Inventory Control Eligius M.T. Hendrix, Rene Wenlin Chen, Haijema, Karin Chung-Li Tseng G.J. Pauls-Worm and Roberto and Shu-Jung Rossi Sunny Yang Buxton, Marwan Khammash, Konstantinos Nikolopoulos and Philip Stern Health Designing Dementia Care With Group Modeling ST.L. Cowan Green Logistics Logistics and Forecasting, Data Mining Environmental Supply Chain & CIM TimeAssessment Of Transport Differentiated Unlocking The Service Parts Potential Of Routes Computer Javier Faulin, Distribution Esteban de Mohsin Jat, Bart Games OR Consultancy Queue Modelling OperationsStochastic /Case Studies DesignInventory Hierarchical Innovation Control SWOT Analysis Queues Interface Forecasting Using General Jason Young and SchumpeIntermittent Morphological Vincent Knight terian EntreDemand Analysis preneurship Steven Prestwich Nasir Hussain, Green Logistics Sustainability And Seafood Banquets, An Utopia? OR Consultancy Queue Modelling /Case Studies A Queueing NHS Provider Model For Landscaping Specialty Nicholas Jones Clinics Paz, Angel Juan and Fernando Lera Altea Lorenzo Arribas, Carmen Cadarso Suárez, Vicente Lustres Pérez and Isabel M. Martínez Silva MacCarthy and Robert Raeside, Luc Mark Robinson, Muyldermans Anusua Singh Roy and Tracey Warner Forecasting, Data Mining & CIM Natural Gas Demand Forecasting Nikolaos Mavroeidis, Stella Androulaki, Vassilios Assimakopoulos and John Psarras 30 Yanto Chandra and Shu-Jung Sunny Yang Bruce Garvey and Tom Ritchey Navid Izady JM Pentland East JM Pentland West Thursday 06 September 2012 Time JM Pentland JM Prestonfield 10:00 10:30 11:00 11:30 FULL CONFERENCE PROGRAMME JM Salisbury JM Duddingston Health Credit Risk Simulation Green Logistics Modelling Corporate Developing The BiThe Dental Governance In Parallelised Objective Workforce In Finance Modelling PollutionSri Lanka Distress Environments Routing Sally Brailsford Zhiyong Li, Richard Boakes Problem and Dileep De Galina Andreeva Silva and Jonathan Crook Credit Risk Aggregating Credit And Mohamed Tadjer Market Risk: Model Choice Bastiaan Verhoef and Gareth Toomey Tolga Bektas, Emrah Demir and Gilbert Laporte Green Logistics Green Tire CLSC Network Design Via EI-99 ST.L. Bonnar ST.L. Cowan Stochastic Forecasting, OR Consultancy Optimisation OperationsInventory Data Mining /Case Studies Dynamic DEA DesignControl & CIM Innovation Managing With Network Interface Discovering Forecasting Projects In An Structure Impact Of A New Order Solvency Of Uncertain Kaoru Tone and Competition Policy Greek Banks Miki Tsutsui World On Capability Rene Christina Fran Ackermann, Haijema Konstantinidou, Susan Howick, Development Vassilis Tom Houghton, Pa Assimakopoulos John Quigley and and Serafeim Lesley Walls Gravanis Stochastic Inventory Control Product Recovery With Kemal Subulan, Substitution Adil BAYKASOGLU and A.Serdar TASAN 12:30 13:00 ST.L. ST.L. ST.L. St.Trinneans Pollock Nelson Refreshments in JMCC ST.L. Brewster Plenary: 10:30 (JM Pentland) - Chris Dent, Durham University - 'Challenges in energy system modelling: data, optimisation and futurology' Health Service Redesign 12:00 JM Holyrood (please note that authors names are listed in the order of Presenter first and then alphabetically) Forecasting, Data Mining & CIM Decision Making In Complex Environments Sarah Marshall and Tom Jose M. Merigo Archibald OR Consultancy Routing Applic's /Case Studies & Transport Prioritising Lower + Asset Upper Bounds Replacement For Joint Liz Archibald Problem Re Routing Lunch 12:30 - 13:30 in JMCC 31 Yan Emma Liu and Shu-Jung Sunny Yang Narges Haghi OperationsDesignInnovation Interface Complex Supply Network Mohsen Jafari Songhori, Alan Smith and Sunny Yang JM Pentland East JM Pentland West Thursday 06 September 2012 continued Time 13:30 JM Pentland FULL CONFERENCE PROGRAMME JM JM JM JM Prestonfield Salisbury Holyrood Duddingston Health Credit Risk Simulation Green Logistics MCDA Stochastic Validation Green Approaches Parametric Of Simulation Intermodal To Time To Models For Freight Prioritisation Default Model Operational TransportVadim In NICE Co ation Brian Reddy Melnitchouk Michael Leyer Yi Qu, Tolga and Sophie Bektas and Julia Meilinger Bennell Health Credit Risk Simulation Green Logistics Location & Default Agile Climate Scheduling Of Probability Of Simulation – Change – A Blood Smes The Way Challenge For Raffaella Collection Forward? MCDA? Calabrese and Frances Sneddon Valerie Belton Sessions 14:00 14:30 Thomas Jeffries and Jesse O'Hanley Silvia Osmetti ST.L. ST.L. ST.L. ST.L. St.Trinneans Pollock Nelson Bonnar Logistics and Forecasting, OR Consultancy Routing Applic's Supply Chain Data Mining /Case Studies & Transport & CIM Reliability Receding Presentations Forecasting And Horizon x4 of 22.5 Resilience In Short Term Approach To minutes each Electricity The Supply Gate Load Demand Chain Allocation 13:30 Eirini-Elisavet Mouhamad Urszula Neuman, Shaker Ali Agha, Theodorou, Haris CB(RN) Risk Jason Atkin and Assessment Edmund Burke Doukas and Robert Van Der MultiMeer and Lesley John Psarras Methodology Walls Diederik Logistics and Forecasting, Routing Applic's Wijnmalen Supply Chain Data Mining & Transport & CIM Two Phase Methods For 13:52 Portfolio Optimization Solving The Understanding Management In Reverse Dynamic Mobile With Logistics PDPTW Workforce Ali Niknejad and Forecasting Penny Holborn, Productivity Rhyd Lewis and Dobrila Petrovic Methods Jonathan Nikolaos Malpass Theodorou, Vassilios 14:14 Assimakopoulos and Georgios Case Studies In Tender Spithourakis Credit Risk UK Smes Through The 'Credit Crunch' Simulation Effects Of Supply Risks: A Simulation Study Meng Ma, Mualla Gonca Galina Andreeva Yunusoglu, and Jake Ansell Hasan Selim and A. Serdar Tasan Evaluation Logistics and Forecasting, David Wrigley Supply Chain Data Mining & CIM Judgemental 14:36 Adjustment Evaluation Of Applying Soft Volatility In An Systems Forecasting Inventory Methodology Models System to Better Bing Xu and Inna Regulation Jamal Kholidasari Ian Mitchell Ouenniche and Aris Syntetos 15:00 (please note that authors names are listed in the order of Presenter first and then alphabetically) Conference closes at 15:00 32 Jonathan Thompson ST.L. Cowan OperationsDesignInnovation Interface Altruistic Entrepreneurship Yanto Chandra, Yan Emma Liu and Shu-Jung Sunny Yang OperationsDesignInnovation Interface Shelf-Space Competition Chia-Wei Kuo and Shu-Jung Sunny Yang Routing Applic's Operations& Transport DesignInnovation Route Interface Generator For Redesign And WSRP Innovate J. Arturo Castillo-Salazar Aging Medical and Dario Landa- Institutes Silva Jiun-Yu Yu ST.L. Brewster JM Pentland East JM Pentland West Optimisation An MPEC Approach To Electricity Trading Strategy Arash Mostajeran Gourtani, David Pozo, Maria Teresa Vespucci and Huifu Xu Optimisation The Stochastic Network Loading Problem Konstantinos Kaparis, Adam Letchford and Stein Wallace PLENARY SESSIONS DR GEOFF ROYSTON Opening plenary: Operational Research for the real world: big questions from a small island 04/09/2012 : 09:30 : Pentland Room in JMCC The address will explore, from an unlikely starting point, some key questions that are important for operational research, seeking to draw lessons from the past, gather ideas from the present and provide messages for the future. It will argue that ‚real world‚ operational research needs to take its subject more seriously and that this requires some developments in both theory and practice. It will conclude by suggesting some implications for education, for research, for practice, for clients and for society. About Geoff Royston: Geoff Royston is former Head of Strategic Analysis and Operational Research in the Department of Health for England and a former Chair of the Government Operational Research Service. He is now an independent analyst and researcher and is currently President of the Operational Research Society, He has had a wide range of activities and responsibilities involving analysis, modelling and research to inform the design, implementation and evaluation of evidence-based policies and programmes in the health field. He has also worked on information and communication technology in the health sector, has been an adviser to the UK communications industry regulator (OFCOM), and led the design and launch of the national telephone advice service NHS Direct. He has been an external examiner for postgraduate courses in operational research and management science, has served on both scientific and medical UK Research Council panels and is a member of the editorial board for the journal Health Care Management Science. He has experience in the international health arena including being a consultant for the World Health Organisation, a member of the expert advisory panel for the HIFA2015 (Healthcare Information for All) initiative and a long standing member of the European Working Group on Operational Research Applied to Health Services. 33 PLENARY SESSIONS MR JOHN HOPES OR and analytics ” an opportunity for growth 04/09/2012 : 17:00 : Pentland Room in JMCC Analytics has gained a high profile in recent years in business and in government, largely driven by the explosion in data volumes in the digital economy (the so called ‚big data‛ issue). It has also been a major focus for the OR Society which has identified it as an opportunity both to expand its offering to its existing members and to reach out to new members in the wider analytics community. This session will set the scene by describing the analytics landscape, the various areas of application, the software market and the reasons why it is seen as a driver of competitive advantage. It will explore the different categories of analytics - descriptive, predictive and prescriptive - and will look at their relationship with OR. It will also summarise some of the key themes that emerged from the OR Society’s recent one day advanced analytics national event, illustrating these themes with case study examples. It will then move on to describe the OR Society’s analytics initiative, the progress that has been made to date and the next steps that are currently being taken to ensure that the Society becomes a key player in promoting analytics and in meeting the needs of the analytics community. It will also summarise the steps that INFORMS has made to support the analytics community in the US. Plenty of time will be allowed for discussion and to receive comments and suggestions on the future direction of this important initiative for the OR Society. About John Hopes: John Hopes is a partner in Ernst & Young’s Business Modelling practice, a team of over 100 people in the UK and over 300 globally. He has over 30 years experience of delivering OR, financial modelling and analytics solutions, both in industry and as a consultant. His career began with Shell where he worked in a variety of roles covering OR, strategic planning and production economics. He then moved into consulting with KPMG where he progressed from senior consultant to partner within the modelling team before becoming leader of the customer management consulting practice. He joined Ernst & Young 11 years ago to lead the Business Modelling practice. John is also currently one of two Vice Presidents of the OR Society. 34 PLENARY SESSIONS MR JASON FIELD Operational Research ” Supporting Defence and Security 05/09/2012 : 09:30 : Pentland Room in JMCC The application of Operational Research (OR) to current day defence and security challenges can allow for a greater understanding, or systems view, of the overall problem space. Use of specific methods and techniques can highlight the impacts and issues associated with potential solutions. Dstl provide support to the Ministry of Defence and many other government departments with the practical use of OR methods ranging from decision support, risk analysis and facilitation through to complex computer modelling and simulation. This input has informed both policy decisions and design of security measures. The work has allowed assessment of the effectiveness of the proposed solutions, accounting for cost, risk and impact to provide the final decision maker with additional useful information. This presentation will focus on the practical application of the OR methods and their place in supporting the decision makers. The presentation will give a background to the variety of techniques used and examples of the output from the work. The presentation will draw from a number of real life case studies in support of defence and security across government from policy support to the design of new security systems. Through these case studies, the presentation will highlight the challenges in the use of OR, its place in the decision making process and consider how best to present the output from studies to maximise the messages for non OR specialists. The presentation will hope to demonstrate that appropriate use of OR can have a significant impact in decision making in defence and security. About Jason Field: Jason has worked for the MoD for the last 23 years as an Operational Researcher. He currently works in the Dstl Programme Office and is the Domain Leader for Security. He has a degree in management science and IT. During his career, Jason has lead teams delivering OR products for MoD, been a MoD OR scrutineer and been a senior analyst responsible for quality of delivery of OR products. Over the last 6 years, Jason has focussed on the challenges of providing S&T and OR support to non MoD government departments. 35 PLENARY SESSIONS DR CHRIS DENT Closing Plenary : Challenges in energy system modelling: data, optimisation and futurology 06/09/2012 : 10:30 : Pentland Room in JMCC Mathematical modelling is used as a means of taking systematic decisions across all aspects of planning and operation of energy systems. This presentation will look in detail at three aspects: Assessment of whether there is sufficient generating capacity installed in a power system in order to support demand with a given level of reliability. Generation adequacy is one of the fundamental aspects of system planning, and impinges on wider society through the dependence of modern life on a reliable electricity supply. However, as with many problems of modelling rare events, directly relevant data are very sparse. This section of the presentation will describe how the first Great Britain statutory adequacy study confronted this issue. Operational scheduling. Power system optimisation problems have a unique structure as power flows are governed by the physical laws of circuit theory. The optimisation routines in current energy system management software will be outlined, and the challenges of operational planning under uncertainty in systems with high capacities of renewable generation, and of system scheduling in future smartgrids with many more small scale controllable generators and demands. The use of models to understand the future evolution of energy systems. Mathematical modelling is increasingly used to understand the evolution of energy systems over timescales of decades. However, due to the complexity of the energy economy, and extreme uncertainty over model inputs such as technology growth, drawing robust conclusions for real energy systems based on the study of models is not straightforward. A survey of current 'futurology' models in use in the UK and of challenges in their development will be presented. About Chris Dent: Chris Dent AFORS is Lecturer in Energy Systems Modelling in the School of Engineering and Computing Sciences at Durham University. Prior to this position, following 7 years in Physics research, he studied for the MSc in OR at Edinburgh University, before moving into energy research in 2007. He has interests across electrical power systems analysis, including risk and reliability modelling, optimisation, renewables integration, economics and network planning. He plays a particular role in developing links between the power systems and mathematical sciences communities, having co-organised the Energy Systems Week at the Isaac Newton Institute for Mathematical Sciences in 2010, coordinated an Energy Systems special issue of the Journal of Risk and Reliability, and founded the annual "Risk and Reliability Modelling of Energy Systems" day at Durham University. He has recently worked with National Grid on the first statutory GB generation capacity adequacy assessment. 36 PRESIDENT’S MEDAL PRESENTATIONS The President’s Medal is awarded for the best practical application of OR submitted to the competition (a wide definition of OR is used). Entries are accepted from both academics and industry-based OR workers and consultants. One of the main qualifications for entry is that the work has been implemented before submission. Criteria for judging include: The level of demonstrable benefit The intellectual and novel content of the solution The likely longevity of the solution The excellence of the OR process Conference delegates attending the President’s Medal plenary session will have the opportunity to express their views as to their preferred candidate. The judges are required to take into account the views of the audience, but are free to arrive at their own decision. Ballot papers will be distributed at the start of the session. Presentations 5/9/2012 : 13:15 : Pentland Room in JMCC From Operational Research to Operational Planning: Shaping the NATO plan for Afghanistan Dr Patrick Rose and Mr Colin Marston, (Dstl) The Defence Science & Technology Laboratory (DSTL) delivers impartial, high-level operational analysis support to the UK Ministry of Defence (MOD) and wider government decision making. DSTL also embeds scientific teams in military Headquarters in the UK as well as operational areas. During 2011, DSTL deployed two teams of civilian volunteers to Afghanistan to directly support military planners in shaping NATO operations. Two major planning conferences in March and November utilised DSTL’s world-leading wargame capability, the Peace Support Operations Model (PSOM), a research based decision-support tool for examining operations and outcomes in complex environments such as Afghanistan. Originally designed to inform future UK strategic planning, PSOM was employed in Afghanistan within a new and bespoke OR process. This simulated the planning, execution and assessment of real world operations, giving senior military and civilian decision makers clear direction and insights that continue to influence and shape the NATO campaign plan in Afghanistan. This paper provides an overview of PSOM and its development into a decision support tool used for Ministry of Defence strategic planning and force structure analysis. It then details the technical, organisational and validation challenges that the DSTL team had to overcome in order to utilise PSOM to support real world military operations. It concludes by offering a brief overview of the conduct of the planning conferences conducted with the ISAF Joint Command during 2011, and gives an assessment of how this type of OR support refined and shaped the cross-Afghanistan campaign plan being executed at the present time of security transition from NATO to local security forces. 37 5/9/2012 : 13:45 : Pentland Room in JMCC Modelling demand and performance in HMRC call centres Stephen O'Donnell, (HM Revenue and Customs) HMRC call centres have received a great deal of bad publicity about their performance. The management of call centre performance relies heavily on queuing theory work first carried out by Erlang in the early 20th century and much elaborated since. But this approach requires copious and detailed forecasting of future demand and resources and while it is excellent for short term performance management HMRC has found it less suitable for medium to long term performance planning. Part of the problem, which any call centre under pressure faces, is it is difficult to establish how much demand the call centre has to handle. The work described here outlines a novel way of measuring the underlying demand faced by a call centre. This measure turns out to be intimately related to the performance of a call centre and by adopting this measure I have been able to build extremely accurate models of medium to long term call centre performance. HMRC adopted this model for managing call centre performance from April 2011 and, with the assistance of the improved understanding of demand and performance the model brings, went from handling 48% of calls in 2010-11 to handling 74% of call in 2011-12. 5/9/2012 : 14:15 : Pentland Room in JMCC Decision Analysis in Nuclear Decommissioning: Developments in Techniques and Stakeholder Engagement Processes Simon Turner and Stephen Wilmott, (Magnox Limited) Magnox Limited have developed new assessment techniques for the purposes of waste management decision-making and stakeholder engagement, that deliver a more open, transparent, and proportionate approach. The development of this new process of decision analysis has also opened the way for improved approaches to BPEO regulation with the Environmental Agencies, and has had, and will continue to have, enormous savings (running into millions of pounds) in terms of both time and ongoing costs for the organisation as we move forward. 38 Analytics Organiser: Jacqui Taylor 04/09/2012 : 13:30 : ST.L. St.Trinneans Code: OR54A751 Tap into the Global Analysis Development Community with R Open Source Software Dr Chris Campbell (Mango Solutions) Operational Research makes use of diverse skills to solve real world problems. The solutions may be composed of elements from many fields, say mathematics, engineering and psychology. No single software environment has access to the range of analytical researchers as the R statistical programming language. Cutting edge statistical methodologies are available from academically sponsored repositories. And economists and psychologists rub shoulders with statisticians and geneticists. This accessible scripting language provides easy access to the work of thousands of specialists. Used it directly to run analyses and create reports, or hide it away as the engine behind a flashy interface. Isn’t it time you learned to speak R? 04/09/2012 : 14:00 : ST.L. St.Trinneans Code: OR54A757 Big Brother is Still Watching, but ... Mrs Jacqui Taylor (FlyingBinary Limited) In our quiet moments, we are all concerned about the level and detail of data collected about each one of us, by the faceless leviathans that make up Government, public services and private commerce. But there are glimmers of hope. Over the last few years, governments have started to open up and share the data they hold (Open Data) and citizens are beginning to understand and care about what this means for them. Accessibility is an issue where visualisation has a key role, turning a torrent of opaque numbers into transparent understanding. Over the next 18 months, there are exciting developments within the Open Data agenda that have the potential to catapult the usability of this medium, impacting all our lives. In this talk by Jacqui Taylor, CEO and co-founder of FlyingBinary, you will hear: - What Open Data has achieved so far Why the current usability is limited - What the next level looks like - Why you should be excited about the changes - How you can be part of shaping the future. 04/09/2012 : 15:30 : ST.L.Pollock Code: OR54A667 Royal Mail Goes To The Olympics! Mr Paul Liddiard (Royal Mail) In 2006 the UK Postal Market was fully liberalised. In April 2012 rivals’ postmen began delivering on the streets of West London. This presentation will look at the methodologies used in forecasting mail volumes and the impact that such competition may have upon those volumes. 04/09/2012 : 16:00 : ST.L.Pollock Code: OR54A756 Enter the Matrix Mr Ian Taylor (FlyingBinary Ltd) You know the scenario. Months of careful analysis, diligent preparation and massive development gives birth to the next campaign / product / service / movie / building / whatever. But the reaction is not as anticipated. Sometimes good, sometimes bad, the voice of "the public" is essential if forward-facing 39 organisations are to reach their full potential. This we call Social Intelligence. But the two worlds "in here" (corporate) and "out there" (public) are as different as the worlds in The Matrix. So can you really determine the voice of the public? Even if you can, does it say anything useful? And if it does - can you act on it? In this session, Ian Taylor, CTO and co-founder of FlyingBinary, will present and demonstrate live capability of visualising internal and external data sets to trigger real action following a marketing campaign. You will see and hear: - What to look for in a Social Intelligence solution - Why visualisation is essential to communicate effectively - How to test if an organisation is ready for Social Intelligence - The real story behind one campaign. Please note the live demonstration is real data, therefore has to be anonymised to protect the original client identity. Just like the movies. 04/09/2012 : 16:30 : ST.L.Pollock Code: OR54A743 An Agile Approach to Knowledge Discovery of Web Log Data Mr Paul Lam (uSwitch.com), Mr Paul Ingles (uSwitch) and Mr Thibaut Sacreste (uSwitch) Web log data contains a wealth of information about online visitors. We have a record of each and every customer interaction for the millions of visitors coming through each month at uSwitch.com. The challenge is to analyse this discrete time series, semi-structured dataset to understand the behaviour of our visitors on a personal level. This talk is a case study of how our data team of three leveraged heterogeneous architecture and agile methodologies to tackle this problem. And we had three months. 40 Community OR and Sustainable Support for Communities Martha Vahl Eliseo Vilalta-Perdomo Organisers: Martha Vahl and Eliseo Vilalta-Perdomo 04/09/2012 : 11:00 : JM Pentland Code: OR54A744 KEYNOTE: Modelling to Support Communities Prof Gerard de Zeeuw (University of Lincoln) and Ms Martha Vahl (Centre for Educational Research & Development) O.R. analysts increasingly are interested in the Third Sector (community and voluntary/charitable sector) as demonstrated by the establishment of the Community O.R. Unit and network in 1988 and the more recent initiation of a Third Sector special interest group. Part of the reasons for this interest is the perceived difficulty of using the tools of O.R., even though O.R. thinking often proves quite effective. This raises methodological and philosophical questions. Of particular relevance is the fact that research issues are not owned by any organisation and hence do not relate to an accepted (acceptable) collective preference. Similar difficulties relate to the emotional engagement of individuals to their community. The problem will be interpreted in terms of Arrow's impossibility theorem. In the paper a number of the methodological and philosophical questions will be discussed and answered. Some examples of research projects will be presented, including an initiative to change the teaching culture in a University. The implications of the answers provided will be discussed in terms of the sustainability of local communities in Lincolnshire. 04/09/2012 : 11:30 : JM Pentland Code: OR54A605 A Communities of Practice (CoP) Approach to Evaluating Expert Support Mr Ricardo Barros (Universidad de los Andes), Prof Gerald Midgley (University of Hull) and Dr Luis Pinzón (Universidad de los Andes) Ulrich’s critical systems heuristics approach encourages participants in planning to question if and how different expert roles can guarantee systemic improvement. Building on this idea, and drawing upon Habermas’s theory of knowledge constitutive interests, Reynolds proposes a generic framework for evaluating expert support. This paper takes the debate one stage further by proposing a conceptual framework to evaluate expert support based on Wenger’s theory of Communities of Practice (CoP). The ideas in CoP are explained, and reflections on their utility for supporting the evaluation of expert support are provided. 04/09/2012 : 13:30 : JM Pentland West) Code: OR54A616 O.R. in the 3rd sector: Where From, Where Now, Where Next? Ms Ruth Kaufman (Independent) It was after a talk in the Community O.R. stream at the 2009 OR Society conference that the idea of an 'O.R. in the Third Sector' initiative first emerged, in a conversation between Andrew Dobson, the presenter, and some of the audience. Three years on, we: have successfully piloted a 'pro bono O.R.' initiative; have established an 'OR in the3rd Sector' special interest group; and, are thinking about what else needs to be done to create a sustainable pro bono O.R. service, and to achieve the aim of expanding the use and effectiveness of O.R. to support charities, social enterprises and other non-profit groups. This presentation 41 will report on the projects that have been undertaken under the initiative, and open up the question of where and how we move forward. 04/09/2012 : 14:00 : JM Pentland West Code: OR54A617 A Framework to Identify and Mitigate Potential Conflicts in the Professional and Research Agendas on Sustainability Dr Eliseo Vilalta-Perdomo (University of Lincoln) The Triple Bottom-Line is now a traditional model to discuss sustainability. It has been helpful to make individuals and organisations aware of the impact of their actions. However, one criticism is that it has been unable to show how to achieve or negotiate a balance between the different dimensions involved (economic, environmental and social). This paper shows different reasons why this may happen. It also proposes where to look to break this current impasse. 04/09/2012 : 14:30 : JM Pentland West Code: OR54A649 Using Problem Structuring Methods (PSMs) to Enrich Project Management in Community Regeneration Mr Michael Charlton (Sheffield Hallam University) Strategic Housing Departments within Local Authorities have the challenging task of coordinating multiple agencies to improve and create new residential communities. The task is usually supported by a traditional project management method employing expert-driven approaches and stakeholder consultation to create masterplans for specific target areas which seek to win approval and support from those affected. This short paper reports some findings from the efforts of one Local Authority in Nottingham to develop new housing for one of its most deprived wards. The paper sets out the general nature of the problems facing the Local Authority, the approach taken, its suitability and weaknesses. An examination of the potential of using selective Problem Structuring Methods (PSMs) in this context is then discussed and some conclusions are drawn about how certain features of PSMs might be deployed to enrich the overall project management approach. Observations are also made on the limitations of PSMs in this particular context. 04/09/2012 : 15:30 : JM Pentland Code: OR54A693 Researching ‘Wicked’ Issues – The ‘Messy’ World Of Community O.R. Ms Jennifer Jackson (University of Lincoln) Health inequalities is considered a ‘wicked problem’ with the understanding of the problem that health inequalities poses subject to interacting causes and uncertain means of resolution of how to act within policy, organisational and academic debate. This paper focuses on a 5-year community operational research project on a healthy living centre partnership within a probation system and the messy issues that it represented. It explores how the notion of the practitioner/researcher entwines with concepts of ‘research’ and action, with the ‘role’ of the O.R. participant continually negotiated and questioned within the process. Ultimately it questions who is the ‘expert’ and what is the problem that needs to be solved? 04/09/2012 : 16:00 : JM Pentland Code: OR54A739 Statistical Analysis of Engineering Tests of TASER Devices in Canadian Police Inventories Mr Fred Cameron, Dr Jack Brimberg (Royal Military College of Canada), Dr Bill Hurley (Royal Military College of Canada), Dr Bill Simms (Royal Military College of Canada) and Mr Xiaofeng Yang (Royal Military College of Canada) Although police agencies have recourse to deadly force under certain conditions, non-lethal weapons have always had a place in the inventory. Notably British police have a proud reputation of often patrolling without lethal weapons entirely. But, like counterparts in other nations, British police agencies have expanded the fielding of weapons like TASERs, a type of conducted energy weapon. With a growing interest in a broader range of non-lethal weapons, TASER International Inc. began developing their line of devices in 42 the 1970s and these have been widely adopted by police services in many nations. In Canada these weapons are the subject of close scrutiny. Public attention to them increased considerably since the death at the Vancouver International Airport of Robert Dziekański, who had been the subject of TASER employment. One aspect of Canadian scrutiny is an ongoing series of engineering tests intended to confirm that weapons in police inventories continue to meet their design specifications. The Royal Military College of Canada was asked to conduct a broad statistical analysis of these test results to determine if there may be patterns that could be used to ensure effective but safe employment in the future. This presentation includes several aspects of the methods that were employed and some general results of the College’s investigation. It includes an outline of how these results may be used to change operating procedures. 43 Credit Risk Management Organiser: Jonathan Crook 06/09/2012 : 08:30 : JM Salisbury Code: OR54A557 Survival Analysis with Unobserved Heterogeneity for Credit Card Loans Prof Jonathan Crook (University of Edinburgh), Dr Tony Bellotti (Imperial College London) and Dr Mindy Leow (University of Edinburgh) Recent literature has presented a number of survival analysis models of the conditional probability that a credit account will default in the next time period. In this paper we advance the literature by presenting comparative empirical survival analysis models that incorporate an unobserved heterogeneity (frailty) term. This term may be interpreted as representing omitted variables or mis-specified functional form. We show, using a large sample of credit card loans, the effect of incorporating such terms under different assumptions about the frailty term and with different functions for the baseline hazard. We find that the parameters of hazard functions are substantially affected by the inclusion of frailty when the baseline function is simplistic, but much less so if the baseline function is very flexible. However we find that the overall predictive performance of hazard models is remarkably robust to the exclusion of frailty terms. 06/09/2012 : 09:00 : JM Salisbury Code: OR54A692 Estimation of Joint Credit Losses based on Poisson Processes and a Suggestion for Basel Accords Dr Fernando Moreira (University of Edinburgh Business School) We suggest a model based on Poisson processes to estimate joint credit losses without the limitations of normality assumptions and compatible with some levels of negative correlation. Idiosyncratic and systematic risks are seen as ‚shocks‛ and defaults are driven by a latent variable (loans’ lifetimes). The method is applied to the calculation of capital to cover unexpected credit losses in financial institutions and its implementation is relatively simple. Simulations show that the general model yields satisfactory estimations of simultaneous credit losses and its application to capital assessment outperforms the current approach (Basel Accords II and III) in some scenarios even when the alternative method is calculated at lower levels of confidence. 06/09/2012 : 09:30 : JM Salisbury Code: OR54A734 Prediction Models of Financial Distress in Chinese Firms from 1991-2011 Dr Anthony Bellotti (Imperial College London) and Dr Xijuan Bellotti (Middlesex University Business School) Prediction models of corporate bankruptcy have been used in North America and Europe since 1968, based on Altman’s Z-score model which uses financial ratios as predictor variables. In these markets, Z-score models are found to remain robust over many years. We build similar predictive models of financial distress in the developing Chinese market based on data covering the period 1991 to 2011, sourced from GTA, and using Special Treatment (ST) status as an indicator of financial distress. We determine the most important financial ratios predictors and compare with models developed in other markets or by previous researchers. In particular, we investigate the robustness of the model over time. We conclude that the effect size of some financial ratios in the model show a large and statistically significant change over time, and 44 forecasting performance deteriorates rapidly over the few years following the initial model build date. These results reflect the changes and developments in the corporate market in China, as capitalism emerged within the communist system, during the 1990’s. 06/09/2012 : 11:30 : JM Salisbury Code: OR54A636 Corporate Governance in Finance Distress: A Case of Chinese Listed Companies Mr Zhiyong Li (Credit Research Centre), Dr Galina Andreeva (Credit Research Centre) and Prof Jonathan Crook (Credit Research Centre) Credit risk pays attention to both corporate bankruptcy and consumer default. Beside financial conditions, corporate governance has impact on corporate performance by influencing its management. But unlike financial ratios as outcomes of performance, behaviours of directors and managers actually are origins of management performance. This paper investigates the influences of some corporate governance measures on the probability of financial distress. Measures of board composition, ownership structure, management compensation and directors’ characteristics are used as variables in prediction, which is very similar in consumer credit scoring. Data from around 2,000 Chinese listed companies provides sources of modelling and there are some significant effects of corporate governance measures on its financial distress. Up to date results will be presented. 06/09/2012 : 12:00 : JM Salisbury Code: OR54A637 Aggregating Credit and Market Risk: The Impact of Model Specification Mr Bastiaan Verhoef (Royal Bank of Scotland) We investigate the effect of model specification on the aggregation of (correlated) market and credit risk. We focus on the functional form linking systematic credit risk drivers to default probabilities. Examples include the normal based probit link function for typical structural models, or the exponential (Poisson) link function for typical reduced form models. We first show analytically how model specification impacts ‘diversification benefits’ for aggregated market and credit risk. The specification effect can lead to Value-atRisk (VaR) reductions in the range of 3% to 47%, particularly at high confidence level VaRs. We also illustrate the effects using a fully calibrated empirical model for US data. The empirical effects corroborate our analytic results. 06/09/2012 : 13:30 : JM Salisbury Code: OR54A546 A Practical Approach to Time to Default Forecasting: Stochastic Parametric Modelling with Macroeconomic Variables and Unobserved Consumer Heterogeneity Dr Vadim Melnitchouk (Metropoliten State University) According to Bellotti & Crook (2007) survival analysis is a competitive alternative to logistic regression when predicting default events. The method has become a model of choice in recent publications. But its complexity makes such a technique unfeasible for practitioners. We apply the stochastic parametric Time to Event method well known in marketing (Fader & Hardie, 2007) to build two models as alternatives to the Cox regression model. Federal Reserve residential mortgage data were used to illustrate the performance of the method. The first model assumes Weibull distribution of Time to Default and gamma distribution of default density across obligors to include unobserved consumer heterogeneity. Monthly number of defaults and time varying covariates like Unemployment and Home Price Index are model inputs. Macroeconomic factors are incorporated into the hazard rate function. Forecast accuracy for out-of-time period is acceptable. According to Bellotti (2010) ‘any credit risk model with macroeconomic variables can’t be expected to capture the direct reason for default like a loss of job, negative equity or a sudden personal crisis such as sickness or divorce’. As an attempt to find an approximate solution to this problem, a latent class Weibull model was used, assuming two segments of obligors with their own fixed but unknown values of shape and scale parameters. We suggest relatively small consumer segment with default hazard increasing over time and relatively large segment with decreasing default hazard. This concept was developed by Hardie & Fader (2001) and used by (Brusilovskiy, 2005) to predict the time of the first home 45 purchase by immigrants. Both models are implemented in MS Excel using ‘Solver’ to get the maximum likelihood estimates of the model parameters. The models can be considered as a simplified aggregate level survival analysis with flexible parameterization. The stochastic parametric Time to Event model has several advantages when comparing with Cox regression and can be used by practitioners to forecast not only Time to Default but Time to delinquency, Time to prepayment and Time to recovery. 06/09/2012 : 14:00 : JM Salisbury Code: OR54A558 Default Prediction of Small and Medium Enterprises by Generalized Additive Models Dr Raffaella Calabrese (University College Dublin) and Dr Silvia Osmetti (Universita' del Sacro Cuore di Milano) We aim at proposing a Generalized Additive Model (GAM) for Small and Medium Enterprises (SMEs). The Generalized Extreme Value regression model (Calabrese and Osmetti, 2011) is extended by replacing the linear predictor with an additive one, defined as the sum of arbitrary smooth functions. In order to focus the attention on the tail of the response curve for values close to one, we consider the quantile function of the generalized extreme value distribution as a link function in a GAM. Thus we propose the Generalized Extreme Value Additive (GEVA) model. To estimate the smooth functions, the local scoring algorithm (Hastie and Tibshirani, 1986) is applied. In credit risk analysis a pivotal topic is the default probability estimation for SMEs. For this reason, we apply the GEVA regression to empirical data on Italian Small and Medium Enterprises (SMEs). On this dataset we compare the performance of the GEVA model with the one of the logistic additive model. The main advantage of the GEVA model is its excellent performance to identify defaults for low default portfolio. Thanks to this characteristic, the drawback of the logistic (additive) regression model in underestimating the default probability (King and Zeng, 2001) is overcome. Finally, the GEVA model is a robust model, unlike the logistic (additive) regression model, if the sample percentage of defaults is different from that in the out-of-sample analysis. 06/09/2012 : 14:30 : JM Salisbury Code: OR54A627 Small and Medium-Sized Businesses through the 'Credit Crunch': Evidence from the UK Miss Meng Ma (The University of Edinburgh), Dr Galina Andreeva (The University of Edinburgh) and Prof Jake Ansell (The University of Edinburgh) Access to credit is a vital condition for SMEs’ survival and development, especially for young start-up enterprises. The credit crisis has put the problem of SME financing to the fore, with governments putting considerable pressure on banks to lend to SMEs. While banks face more uncertainty in SMEs lending, for falling in credit supplying and considerable economic shifts make SMEs’ performance distinct from normal condition. This presentation looks at the credit risk modelling of the UK SMEs via logistic regression applied to a large proprietary dataset, and explores the behaviour of small businesses during the credit crisis and beyond (2007-2010) across industrial sectors, geographical regions, financial situation and for start-ups versus mature businesses. 46 Forecasting, Data Mining & Computationally Intensive Methods Fotios Petropoulos John Crocker Organisers: Fotios Petropoulos and John Crocker 04/09/2012 : 11:00 : ST.L.Pollock Code: OR54A585 KEYNOTE: How Difficult Is It To Forecast The Worlds’ Most Successful Products? The Case of Apple iPhone Dr Fotios Petropoulos (Lancaster University Management School), Prof Vassilios Assimakopoulos (National Technical University of Athens), Mrs Akrivi Litsa (National Technical University of Athens), Prof Konstantinos Nikolopoulos (Bangor University) and Dr Yiannis Polychronakis (University of Salford) Apple is the largest publicly traded company in the world by market capitalization, as well as the largest technology company in the world by revenue and profit. Accurately forecasting unit sales for this giant is of major importance, as even the smallest of the errors will have a huge impact in terms of sales management and revenues. The current research investigates extrapolation alternatives regarding iPhone unit sales, probably the most influential product of the corporation. Simple and straightforward time series techniques (naïve, exponential smoothing methods, theta model) are competing with judgmental approaches (unaided judgment, interaction groups) performed by three groups of participants in terms of expertise (non-experts, semi-experts, experts). Two subgroups (for each level of expertise) were formed, regarding the amount of information provided to participants. The performance of the alternatives was measured in terms of bias and accuracy, using simple error and absolute percentage error. Results indicate that judgmental approaches produced greater (in volume) forecasts than quantitative methods, while focusing effect tends to lead to cognitive bias when forecasting future outcomes. As far as performance is concerned, a combination of time series extrapolation techniques with judgmental forecasts derived from expert groups with limited knowledge produces the smallest forecasting errors. 04/09/2012 : 11:30 : ST.L.Pollock Code: OR54A612 Demand Forecast of Cakes Based on the Kalman Filter Considering Media Effects Mr Ryunosuke Usami and Prof Ichie Watanabe (Japan Industrial Management Association) The Kalman filter, also known as linear estimation, is an algorithm which uses a series of measurements, observed over time, containing noise and other inaccuracies. The Kalman filter operates recursively on streams of noisy input data to produce a statistically optimal estimate of the system state. Usually, when making a long-term prediction using the Kalman filter, a longer-term prediction leads to a larger prediction error, and the predictive values become more ambiguous. In this study, we examine the effectiveness of the Kalman filter for high-precision long-term forecasts. To diminish the error, we propose a model considering factors that have effects on market demand, such as day effects, season effects, and event effects. We focus especially on events in this study, highlighting typical events in Japan. First, we take into account the type of event. For instance, before Valentine's Day, the demand for candies increases as the day approaches and returns to normal immediately after Valentine’s Day. On the other hand, if a product is featured in the media, such as on TV or in magazines, awareness of the product will be heightened by these media. In this case, the demand increases over several days and then decreases gradually. In this study, we combine the Kalman filter with event effects and forecast the sales of cakes, considering various factors of the demand. 47 04/09/2012 : 12:00 : ST.L.Pollock Code: OR54A683 Long-term Forecasting of Sustainable Energy's Share in Total Energy Consuption using Time-series Models: The Case of European Union Miss Nikoletta Koleri (National Technical University of Athens), Prof Dimitrios Askounis (National Technical University of Athens) and Miss Charikleia Karakosta (National Technical University of Athens) Renewable energy use is growing at a much faster pace than the rest of the economy in Europe. This fact along with the necessity for the implementation of the Kyoto Protocol and the world-wide introduction of tradable green certificates have led to a remarkable re-evaluation of the renewable energy sector by politicians and financing institutions. Therefore, every member state of the European Union has set specific targets concerning the amount of energy produced by renewables as a share of its overall energy consumption. The main purpose of this paper is to present a long-term forecast of the intrusion of renewable energy in gross energy consumption of every E.U. country and to evaluate whether the targets set,will be met. To this end, time-series forecasting models were used upon annual data, classified per member-state. More specifically, the forecasting succeeded through the utilisation of exponential smoothing models, such as SES, Holt and Damped models, the linear regression model and THETA model along with the naive model. The derived projections were evaluated and compared with those presented by official or scientific institutions. What is of most importance in this approach, is that time-series models are used also to quantify the policy each country of the European Union implements towards sustainability. 04/09/2012 : 13:30 : ST.L.Pollock Code: OR54A573 Forecasting the Success of Policy Implementation Strategies with Structured Analogies, Delphi and Interaction Groups Prof Konstantinos Nikolopoulos (Bangor University), Prof Vassilis Asimakopoulos (National Technical University of Athens), Miss Akrivi Litsa (National Technical University of Athens) and Dr Fotios Petropoulos (Lancaster University Management School) Forecasting special events such as economic crises, marketing promotions and epidemics can be very challenging, not only due to the nature of those events but also the limited amount, if any, of available historic data from which a reference base can be built. This study uses one such event, policy implementation, in order to evaluate the relative performance of various forecasting methods. Earlier research suggested that individual experts using Structured Analogies could successfully forecast in such contexts, and the present study builds on this by evaluating the use of Structured Analogies blended with group judgmental forecasting techniques, namely Delphi and Interaction Groups. Two real policy implementation cases in an EU country were used to test our methods. Empirical evidence from this study reveals an average improvement of 5% when using Structured Analogies. This improvement is greater when the level of expertise increases, as well as when more and more-relevant analogies are used. Furthermore, if pooling of analogies is ‘forced’ through an Interaction Group, then the actual forecasting improvement could be as high as 40%. 04/09/2012 : 14:00 : ST.L.Pollock Code: OR54A570 Conditional Kernel Estimation for Wind Power Quantile Forecasting Dr Jooyoung Jeon (Strathclyde University) and Prof James Taylor (University of Oxford) There is strong appeal to developing a method that is able to estimate tail quantiles or prediction intervals of wind power. Estimates of tail quantiles, such as 5% and 95%, enable transmission operators and wind farm operators to adapt their strategies by taking preventive actions on higher spinning reserve and machine cut-off. Such quantile estimates also provide useful information to support trading based on future production. Furthermore, recent studies have noted that a quantile forecast is an optimal point forecast in situations where there is an asymmetric cost function. In this study, we consider the estimation of wind power tail quantiles, conditional on the probability distributions for wind speed and direction. Our approach employs conditional kernel density estimation, with the new feature that the procedure is optimised towards estimation of the specific quantile of interest. We term this new method 'conditional kernel quantile' (CKQ) estimation. We compare its performance with the quantiles derived from a more standard 48 conditional kernel density (CKD) method, and simpler benchmark methods. We also investigate the benefit of a univariate conditional kernel approach, which involves conditioning on lagged values of wind power. 05/09/2012 : 08:30 : ST.L.Pollock Code: OR54A750 Predictive Maintenance Modelling with Dynamic Bayesian networks Dr Ken McNaught (Cranfield University) and Dr Adam Zagorecki (Cranfield University) We consider the application of dynamic Bayesian networks to the prognostic modelling of equipment in order to better inform maintenance decision-making. We begin with a brief overview of Bayesian networks and their application to reliability modelling. An example DBN is then presented to illustrate the main features of the approach. The on-going development of a tool to semi-automate DBNs for prognostic and maintenance modelling will be discussed. 05/09/2012 : 11:00 : ST.L.Pollock Code: OR54A587 A Comparison of Relative Accuracy Measures Used in Prediction Dr Christopher Tofallis (University of Hertfordshire) Despite extensive research comparing prediction methods and numerous forecasting competitions, there seems to be little agreement on a measure of relative accuracy. Moreover there is often an inconsistency in that these accuracy measures are used for assessment but not in calibrating the prediction method. We shall present results where various relative accuracy measures are used in the model building process. We then compare the performance including consideration of bias, i.e. tendency for over-prediction or underprediction. 05/09/2012 : 11:30 : ST.L.Pollock Code: OR54A640 Classifying Students’ Performance Through Mining Pattern Behaviour in Program Writing Ms Norhafiza Ahmad (Universiti Malaysia Pahang) In programming, students tend to have their own habits and behaviours while writing and coding a computer program. To complete and execute the program codes, they have to use the keyboard and mouse, actions that are called keystrokes events. This paper will describe and discuss the techniques to mine the behaviour of keystrokes event in program writing from the log files that captured in the laboratory session. To produce the behaviour pattern, the text mining techniques will be applied to analyse the data by classifying students’ performance on program execution and the sequential pattern mining will be approached as the algorithm to mine the frequently occurring pattern related to time consumes and sequences of events while writing the programs. 05/09/2012 : 12:00 : ST.L.Pollock Code: OR54A740 Fast Anomaly Detection in Spatial Point Patterns Using Nearest Neighbours Dr Dafydd Evans (Cardiff University) We present novel exploratory techniques for detecting anomalies in spatial point patterns. Our approach is based on the nearest-neighbour relations between the sample points, which render our methods both computationally efficient and adaptive to smooth variations in the underlying density field. First, we present the main ideas that form the basis of our methods. Under the null hypothesis of local uniformity (or complete spatial randomness), the distance between a sample point and its nearest neighbour should be approximately equal to the distance between its nearest neighbour and the nearest neighbour of its nearest neighbour. If this fails to hold, we assert the existence of an abrupt change in the density field between the sample point and its nearest neighbour, relative to the spatial resolution afforded by the data. The density field will be represented as an inhomogeneous Poisson point process: in spatially homogeneous regions, it is known that the nearest neighbour volume of a sample point (defined to be the volume of the ball centred at the sample point with the nearest neighbour of that point on its boundary) has exponential distribution with inverse-scale parameter equal to the local intensity. Under the null hypothesis of local uniformity, this 49 allows us to derive an exact expression, conditioned on the nearest-neighbour volume of a sample point, for the distribution of the nearest-neighbour volume of the nearest neighbour of that sample point. The analysis extends naturally to the case of second-nearest neighbours, third-nearest neighbours and so on, with the exponential distribution replaced by a gamma distribution with shape parameter equal to the number of nearest neighbours under consideration. Increasing the number of neighbours decreases the variance of our estimators, but also reduces the spatial resolution of detection algorithms. Next, we present two sample statistics based on these ideas. The first is defined in terms of the difference between nearest neighbour volumes, and serves to estimate the directional derivative of the density field along the vector from a sample point to its nearest neighbour. The second is defined in terms of the logarithmic ratio of nearest neighbour volumes, and is similarly intended to quantify the local divergence of the density field. Using the exact distribution of these statistics under the null hypothesis of local uniformity, we can identify a set of critical edges in the nearest neighbour graph of the spatial point pattern. Finally, we test our algorithms on a set of synthetic point patterns containing anomalies of varying intensity, and discuss some possible applications. 06/09/2012 : 08:30 : ST.L.Pollock Code: OR54A550 Modelling and Forecasting Pharmaceutical Life Cycles: Branded versus Generic Miss Samantha Buxton (Bangor University), Dr Marwan Khammash (University of Sussex), Prof Konstantinos Nikolopoulos (Bangor Business School) and Prof Philip Stern (Loughborough University) This paper discusses modelling and forecasting pharmaceutical life cycles specifically focusing on those where sales of branded drugs decline as the prescription rate of the generic drugs increases. This is important as there are currently very few new drugs being produced; pharmaceutical companies are looking to extend the market life span of drugs already in the marketplace, and the market in the UK alone is currently worth $19843 million. Losses from pharmaceutical mismanagement can be significant and dramatic. A Bass diffusion and repeat purchase diffusion model are used to model and forecast these life cycles along with a simple naïve model, exponential smoothing and moving average models. The results show that the RPDM diffusion model is the most accurate in both modelling and forecasting branded versus generic pharmaceutical life cycles. 06/09/2012 : 09:00 : ST.L.Pollock Code: OR54A565 Unlocking the Potential of Computer Games Prof Robert Raeside (Edinburgh Napier University), Mr Mark Robinson (GamesAnalytics), Miss Anusua Singh Roy (GamesAnlytics) and Ms Tracey Warner (GamesAnalytics) The computer games industry has grown rapidly and especially recently with the widespread uptake of social network media such as facebook, and this has led to a substantial positive impact on the economy. In 2008 it is estimated that there were around 10,000 games developers in the UK contributing some £400 million to GDP and £130 million to the Exchequer. But competition is fierce and subject to the whims of fashion. For games designers key performance indicators are the retention of players to the game, accumulation of micro purchases and the number of new players invited to join the game by current players. These should be optimised, which necessitates understanding player behaviour and the decisions they make. This, however, is difficult. There can be many thousands, if not millions, of players with many different goals. Game events are identified for each player and the time of each event recorded so this leads to very large data matrices containing many millions of cells. A way to gain insight is to apply traditional database segmentation strategies to organise players into more homogeneous groups and then to apply standard statistical decision making tools such as logistic regression to determine the likelihood of an action and survival analysis methods to predict retention times. In this paper we report on an approach to apply these methods in a structured way that allows the vast quantities of data to be handled. This approach is having some success in unlocking the potential of games. 50 06/09/2012 : 09:30 : ST.L.Pollock Code: OR54A584 Natural Gas Demand Forecasting: Analysis and Modelling Mr Nikolaos Mavroeidis (National Technical University of Athens), Ms Stella Androulaki (National Technical University of Athens), Prof Vassilios Assimakopoulos (National Technical University of Athens) and Prof John Psarras (National Technical University of Athens) In the last decade natural gas shows particularly great penetration in the energy market and ever increasing shares in relation to other fuels. This is mainly because natural gas is a cleaner fuel with lower price ranges than oil. In Greece the use of natural gas is still relatively immature, as gas production only supplies a low percentage of the domestic energy system, although there are great investments that have taken place. For these reasons, the process of forecasting natural gas consumption is particularly important to distribution companies. In this study, we estimate natural gas consumption in Greece for short, medium and long term horizons, where weekly, monthly, and yearly time series of data are studied respectively. Moreover, the available empirical data are explored in terms of distribution channels, geographical areas and industrial/residential users. Forecasts are made based on time series techniques such as exponential smoothing methods and Theta model as well as combinations of the above. In addition, we explore an explanatory model for long term gas consumption, using as independent variables GDP, year and mean price. Lastly, we evaluate the alternatives’ performance via a series of accuracy metrics. 06/09/2012 : 11:30 : ST.L.Pollock Code: OR54A540 Forecasting Solvency of Greek Banks: A Neural Network Approach Mrs Christina Konstantinidou (NTUA), Prof Vassilis Assimakopoulos (NTUA) and Mr Serafeim Gravanis (Undergraduate Student) The dramatic condition in the Greek economy during the last years and the interaction of the economic situation with the banking sector led to the need of bankruptcy prediction models in order to find the indicators that influence the banks’ solvency. As the financial situation of the banking sector is an important prerequisite for the economy, we tried to develop an ANN model in order to forecast the behaviour of five major Greek banks. As independent variables related to the bank profitability we considered treasury bonds, interbank deposits and total deposits, as well as the macroeconomic parameters, gross national product, gross domestic product, wholesale price index, consumer price index, net external debt, euro exchange rate and net international reserves. The dependent variables were the ratios of profits to asset, equity to asset and non-performing loans to total loans. The forecasts derived from the proposed model were compared with those of standard time series methods (Naive, SES, Holt etc) as well as regression models. The results indicate that neural networks provide more accurate forecasts for banking failure. 06/09/2012 : 12:00 : ST.L.Pollock Code: OR54A700 Decision Making in Complex Environments with Fuzzy Generalized Aggregation Operators Dr Jose M. Merigo (University of Manchester) We introduce a new approach for dealing with complex environments where the available information is represented with fuzzy numbers. We present the fuzzy generalized unified aggregation operator (FGUAO). It is an aggregation operator that unifies a wide range of aggregation by adding a new aggregation that includes all the particular sub aggregations and considering the degree of importance that each case has in the analysis and in an uncertain environment where the information is given in the form of fuzzy numbers. The main advantage of this approach is that we can represent different perspectives of the same problem considering different sources of information. We also study its applicability and we see that it is very broad because all the previous studies that use an aggregation operator such as the weighted average, the probabilistic aggregation or the OWA operator can be revised and extended with this new approach. We focus on a decision making problem regarding the European Union (EU) decisions. For doing so, we use a multi-person aggregation process obtaining the multi-person ” FGUAO (MP-FGUAO) operator. Its main advantage is that it permits to deal with the opinion of several experts in the analysis. We analyse the EU decision concerning the rate of the public deficit that an economy should have in order to reach the 51 convergence criteria required for entering the euro zone. We see that each particular type of FGUAO operator may lead to a different decision. 06/09/2012 : 13:30 : ST.L.Pollock Code: OR54A583 Forecasting Short Term Electricity Load Demand: A Case Study Miss Eirini-Elisavet Theodorou (National Technical University of Athens), Dr Haris Doukas (National Technical University of Athens) and Prof John Psarras (National Technical University of Athens) The purpose of the current research is efficient short term electricity demand forecasting. Electricity demand patterns are characterised by two seasonal cycles (weekly and hourly) doubled with variances regarding special days, such as holidays and strikes, as well as extreme weather conditions. Empirical data used for this study are hourly Greek electricity demand observations (in MWh) of Low Voltage for 32 months. We investigate the performance of a two-step forecasting methodology. Firstly, short term extrapolation of data is performed by a batch of alternative techniques (Naive, SES, Holt, Damped, Taylor’s expansion of HoltWinters’ method and Theta model), while classical decomposition is applied where necessary. Then, the produced forecasts are adjusted so as to incorporate impact of special days. The accuracy of the proposed approach is tested by rolling evaluation (sliding simulation) of 24-hours ahead forecasts for 8 months. Preliminary results indicate superiority of simpler extrapolation techniques, while underlying the need for a posteriori adjustments. 06/09/2012 : 14:00 : ST.L.Pollock Code: OR54A590 Investment Portfolio Management with Trend Forecasting Methods Mr Nikolaos Theodorou (National Technical University of Athens), Prof Vassilios Assimakopoulos (National Technical University of Athens) and Mr Georgios Spithourakis (National Technical University of Athens) This study addresses the problem of portfolio management in modern financial markets. Over the past decade various trading algorithms have become a major factor in traders’ investment decisions. In two of today’s greatest stock exchanges (NYSE and LSE) algorithmic trading is used for roughly 50% of everyday trading. The purpose of this study is to examine whether statistical forecasting methods can be applied in portfolio management. The data used are daily closing prices of the SP500 and the FTSE All-Share components as well as the whole of the Athens Stock Exchange listed companies, covering a five-year period. Using the last two years for simulating the stock selection process, the results of three forecasting methods (Holt, Damped and Theta) lead to the development and processing of an investment portfolio. The aim is to define a robust method, suitable for use in all markets. The variables are the forecasting horizon, the data used for producing the forecast, the size of the portfolio and the way the stocks are selected (by highest forecasted profit, percentage growth etc.). The investment portfolio created with the proposed methods leads to profit up to 40% in two years in both the New York and London stock markets. 06/09/2012 : 14:30 : ST.L.Pollock Code: OR54A686 A Framework for the Performance Evaluation of Volatility Forecasting Models Dr Bing Xu (RGU) and Dr Jamal Ouenniche (University of Edinburgh) Xu and Ouenniche (2012) proposed a multidimensional framework for the performance evaluation of competing models of forecasting crude oil prices' volatility. They used the radial super-efficiency Data Envelopment Analysis (DEA) model to obtain a single ranking that takes account of several criteria. However, by design, radial super-efficiency models cannot deal with negative or zero values in the data, on one hand, and radial measures of efficiency do not take account of the existence of slacks, on the other hand. In addition, under the variable returns-to-scale (VRS) assumption, an input-oriented efficiency score can be different from an output-oriented efficiency score, which would lead to different rankings. Furthermore, radial super-efficiency models might face infeasibility issues. In this paper, we propose a more robust framework that is based on non-radial slack-based super-efficiency DEA models to overcome the above mentioned deficiencies. 52 Government O.R. Service Vivienne Raven Tony O’Connor Organisers: Tony O'Connor and Vivienne Raven 04/09/2012 : 18:00-18:30 : JM Pentland Tony O'Connor (Dept. of Health) and Vivienne Raven (HM Revenue & Customs) The Government O. R. Service, GORS, has several methods of recruitment including the Analytical Fast Stream, central recruitment and job specific vacancies in Departments. In this talk, Vivienne Raven, will summarise the different processes and help you, as academics, advise your students on the best path to take. Open to all, but recommended for academics who have students looking for work in Government O.R. 53 Green Logistics Organiser: Tolga Bektas 04/09/2012 : 15:30 : ST.L.Brewster Code: OR54A532 Combining Electric and Gas Vehicles in Routing Problems: The VRP with Multiple Driving Ranges Dr Angel A. Juan (IN3-Open University of Catalonia), Dr Tolga Bektas (University of Southampton), Prof Javier Faulin (Public University of Navarre) and Dr Jarrod Goentzel (Massachusetts Institute of Technology) We introduce the Vehicle Routing Problem with Multiple Driving Ranges (VRPMDR) and propose a methodology to solve it. The VRPMDR is relevant given the emergence of fleets that add new electric and hybrid-electric vehicles to the common diesel fleet. While diesel and hybrid-electric vehicles can cover any distance, the maximum distance an electric vehicle can cover is limited by the capacity of its batteries. This new challenge has not been properly addressed so far. The paper proposes a multi-round method for the VRPMDR that iteratively constructs a solution using vehicles with different driving ranges. At each round, we use a biased randomisation of a classical heuristic to obtain a ‘good’ partial solution. After the last round, all partial solutions are merged into a global solution. This parameter-less method is able to offer various fleet configurations in a reasonable time period. Some experiments contribute to illustrate and validate our approach. 04/09/2012 : 16:00 : ST.L.Brewster Code: OR54A663 Grocery Deliveries from Depot to Store: Punctual or Green? Dr Daniel Black (University of Edinburgh Business School) and Prof Richard Eglese (Lancaster University Management School) One recent development within the Vehicle Routing Problem (VRP) literature is the incorporation of time dependent travel times due to, for example, road congestion into models (the TD-VRP). Other work has looked at how models can include information on fuel usage or greenhouse gas emissions. This study investigates the relationship between these factors and the punctuality of deliveries. Inevitably there is a trade-off between punctuality and emissions. This is of particular practical interest when these factors have different priorities for different members of the supply chain. The daily delivery schedule of a regional supermarket is used as a case study. 05/09/2012 : 08:30 : JM Holyrood Code: OR54A545 The Time-Dependent Pollution Routing Problem (TDPRP) Ms Anna Franceschetti (Technische Universiteit Eindhoven), Prof Gilbert Laporte (Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (CIRRELT)), Prof Tom van Woensel (Technische Universiteit Eindhoven), Dr Tolga Bektas (University of Southampton ) and Dr Dorothee Honhon (Technische Universiteit Eindhoven) This talk will present an extension to the Pollution Routing Problem (PRP), namely the Time-Dependent Pollution Routing Problem (TDPRP), which takes into account time-dependent speeds in cases where there is traffic congestion. More formally stated, the TDPRP is concerned with routing a homogenous fleet of 54 vehicles to serve a set of customers with known demand. The overall objective is to determine the set of routes for the fleet of vehicles so as to minimise a cost function that accounts for greenhouse gas emissions, fuel consumption, driver travel times and their costs. The problem captures the complexity of the real world traffic conditions by considering time-dependent travel times between nodes and allows for idle waiting at customer nodes which may result in avoiding congestion. The main idea of this study is to focus on insights drawn from small scale instances and to derive analytical results. More specifically, the talk will describe how allowing for idle waiting at the depot and client nodes can be used to mitigate the impact of congestion and to optimise the trade-off between the different cost components. Computational results obtained on realistic instances and insights will be presented. 05/09/2012 : 09:00 : JM Holyrood Code: OR54A547 Fuel Emissions Optimisation in Vehicle Routing Problems with Time-Varying Speeds Prof Richard Eglese (Lancaster University) and Ms Jiani Qian (Lancaster University) The problem considered in this paper is to produce routes and schedules for a fleet of delivery vehicles that minimise the fuel emissions in a road network where speeds depend on time. A model is constructed where the carbon dioxide emissions depend on the speed of the vehicle. Data on traffic speeds at different times of the day on different roads is used to provide maximum speeds due to congestion for the vehicles. However in the model, the speed of the vehicle along each road in its path is treated as a decision variable. A column generation based tabu search algorithm is designed and presented to solve Vehicle Routing Problems where the vehicle routes are limited by the capacities of the vehicles and time constraints on the total length of the route. The best route and schedule for each vehicle is evaluated by a heuristic approach. The method is tested with real traffic data from a London road network. The results indicate that about 6-7% savings in fuel emissions can be achieved when the objective is minimising pollution instead of time. The analysis also shows that most of the fuel emissions reduction is able to be attained in practice by choosing an appropriate route and travelling as fast as is allowed by the traffic conditions up to a preferred speed. 06/09/2012 : 08:30 : JM Duddingston Code: OR54A542 Dynamic Collection Scheduling using Vehicle Routing and Remote Asset Monitoring: A Case Study with Oxfam Mr Fraser McLeod (University of Southampton), Dr Tolga Bektas (University of Southampton), Dr Tom Cherrett (University of Southampton) and Dr Gunes Erdogan (University of Southampton) Of increasing interest in the logistics sector is the way in which ‘remote monitoring’ of assets (e.g., vehicles, trailers, containers, individual parts, staff) could aid more sustainable and efficient collection/delivery scheduling. One area where this concept could realise benefits in the logistics sector is managing the removal of domestic and commercial wastes and recyclate, which is a significant source of logistics activity in the UK, particularly given the often separate tasks of ‘collection’ and ‘disposal’, and the current emphasis on ‘re-use and recycling’. Many collection operations run on the traditional fixed round, fixed time service, where commercial and domestic customers will have their bins serviced at agreed intervals. This talk will present a case study with Oxfam where a series of their donation banks are being equipped with remote monitoring technology to report fill rates at set intervals. These data are then used to investigate how the current fixed donation bank service schedules could be better optimised using real time feedback on donation rates. The talk will describe a ‘time-fluid’ approach to waste collection by using remote fill monitoring technology to enable dynamic optimisation of collection schedules to better meet the true servicing needs of customers. The approach is based on modelling the system as a Heterogeneous Selective Vehicle Routing Problem with Time Windows and additional constraints on Driver Working Hours. The talk will present computational results obtained using the real-life data provided by Oxfam through the proposed approach. 55 06/09/2012 : 09:00 : JM Duddingston Code: OR54A549 Environmental Assessment of Transport Routes in the Pyrenees to Make Distribution Decisions in Navarre (Spain) Prof Javier Faulin (Public University of Navarre), Mr Esteban de Paz (Public University of Navarre), Dr Angel Juan (Open University of Catalonia) and Dr Fernando Lera (Public University of Navarre) Our contribution summarises an environmental assessment of the main routes traversing the Pyrenees in Navarre (Spain) by means of multi-criteria techniques. We focus our study on finding the route sections most impacted by air and noise pollution. A Geographic Information System (GIS) was used to obtain the information that encompasses the section routes and to perform the routes study. Thus, multi-criteria methods (Analytic Hierarchy Process-AHP) (Saaty, 2001) were implemented in order to assess different alternatives according to the considered environmental criteria. The selected criteria are based on social, economic and natural factors (Weintraub et al., 2007). Due to its complexity, we can consider that this decision problem has a finite set of possible solutions: the five alternative routes crossing the Pyrenees. For each criterion considered, we obtain the characteristics of each route, through GIS analysis, and using an AHP method we obtain the solution that best meets the intended purpose of minimising environmental impact considering the aforementioned factors. Among the different impacts produced by road transportation which are considered in the construction of our AHP model we mention the next ones (Sinha and Labi, 2007): a) Rise of pollution atmospheric level because of the traffic; b) Decrease of the sound comfort cause of the traffic; c) Rise of the forest fire risks; d) Reduction in the environmental value of some appreciated countryside area; and e) Affection to the protected areas. Moreover, a DSS (Decision Support System) was built to develop a spatial interface to make easier the interpretation of solutions. Thus, a simulation of size 5,000 was randomly made, tuning carefully the model weights belonging to factors levels, in order to have a wider spectrum of potential solutions for the problem. Finally, the AHP model was enriched using a Delphi method with a feedback procedure. 06/09/2012 : 09:30 : JM Duddingston Code: OR54A614 Sustainability and Seafood Banquets, Just a Utopia? Miss Altea Lorenzo Arribas (Universidad de Santiago de Compostela), Dr Carmen Cadarso Suárez (Universidad de Santiago de Compostela), Dr Vicente Lustres Pérez (Universidad de Santiago de Compostela) and Mrs Isabel M. Martínez Silva (Universidad de Santiago de Compostela) Seafood trade has an enormous socio-economic interest in Galicia (NW Spain). In a context of global over exploitation of marine natural resources, the future of this sector lies in sustainable approaches that will guarantee the reproductive success of the exploited populations as well as the profitability of catch levels. Sea urchin (Paracentrotus lividus, Lamarck, 1816) is one of the seafood species caught for human consumption in the Galician coast and is mainly exported to countries like France where it is highly valued commercially. Average price, capture numbers and sales figures are highly variable and dependent on economic, geographical, political and environmental exogenous factors. This study intends to understand the current model of exploitation of these marine resources and attempts to propose a sustainable framework for its long-term supply-chain management. Institutional time-series data provided by the different fish markets regarding prices, sales and captures, as well as Spanish and European fisheries regulations information and data on distribution and reproductive cycle of this species have been analysed in order to account for both the financial and green factors. An integrative model including current strategies, predictions and potential environmental-based policies has been built as a means to show the sector´s potential for contributing to sustainable development. This model succeeds in portraying the tradeoff that exists between reproduction and commercialisation of Sea urchin P. lividus, and optimises the benefits for the involved stakeholders. Further work will focus on the exploitation of other species with the aim of achieving a unified framework for the sustainability of the Galician seafood market. 56 06/09/2012 : 11:30 : JM Duddingston Code: OR54A539 The Bi-Objective Pollution-Routing Problem Dr Tolga Bektas (University of Southampton), Mr Emrah Demir (University of Southampton) and Prof Gilbert Laporte (HEC Montréal) This paper studies a bi-objective Pollution-Routing Problem, an extension of the Vehicle Routing Problem with Time Windows, where one objective is to minimise fuel consumption, and the other total driving time. A solution method is described where an Adaptive Large Neighbourhood Search algorithm, combined with a specialised speed optimisation algorithm, is used as a search engine for the application of four a posteriori methods to generate Pareto-optimal solutions to the problem. Results of extensive computational experimentation show the efficiency of the proposed approaches. 06/09/2012 : 12:00 : JM Duddingston Code: OR54A709 An Application of Eco-Indicator 99 Methodology for Designing a Dynamic Green Tyre Closed-Loop Supply Chain Network with Multiple Recovery Options Mr Kemal Subulan (Dokuz Eylül University), Prof Adil BAYKASOGLU (Dokuz Eylül University) and Dr A.Serdar TASAN (Dokuz Eylül University) More than a billion brand new tyres are manufactured by approximately 500 producers all over the world. Similarly, nearly a billion scrap tyres are disposed of every year, which results in health hazards and environmental problems. Since such high amounts of used tyres are disposed of using traditional methods that are not environmental friendly, several recovery alternatives have become vital for the last decade. Managing end-of-life tyres effectively and balancing forward and reverse flows in the value chain are challenging tasks in the tyre industry. Designing an economically and ecologically optimised closed-loop supply chain (CLSC) network is therefore a pre-requisite for tyre producers for increased environmental responsibility and sustainable development. Based on this motivation, this paper presents a holistic modelling approach for a tire CLSC using mixed integer linear programming. Alternative recovery options such as re-treading (remanufacturing), recycling and energy recovery are jointly considered in the proposed model. The main objective of this paper is to develop a multi-objective, multi-echelon, multi-product and multi-period logistics network design model while taking environmental issues into account. For quantifying the environmental impact of the CLSC, a life cycle assessment based and damage oriented method (Ecoindicator 99) is utilised. The model is applied to an illustrative example based on the tyre industry in the Aegean region of Turkey and an interactive fuzzy goal programming approach is described for solving the proposed fuzzy multiple-objective network optimisation model. Solutions to the model were obtained by using the optimisation suite ILOG OPL Studio version 6.3. 06/09/2012 : 13:30 : JM Duddingston Code: OR54A615 Green Service Network Design for Intermodal Freight Transportation Miss Yi Qu (University of Southampton), Dr Tolga Bektas (University of Southampton) and Prof Julia Bennell (University of Southampton) Intermodal freight transportation concerns shipping of commodities from their origins to destinations using a combination of different transport modes. Traditional logistics models have concentrated on minimising operational costs by appropriately determining the service network and the routing of commodities. In recent years, the consideration of issues connected to greenhouse gas emissions has led to new models. In this paper, an intermodal transportation problem is described with an explicit consideration of greenhouse gas emissions. A model is proposed which extends the traditional service network design models by taking costs of emissions into account. The objective is to minimise the total costs in an intermodal network, including the capital costs, operational costs and the cost of greenhouse gas emissions. The decisions to be made comprise: (i) the selection of routes and transport modes; and (ii) the flow distribution through the selected route and mode. In order to demonstrate the application and guide the development of our model, a real-life multi-commodity capacitated intermodal transportation network drawn from the UK is presented. By considering various scenarios of costs, results of a set of computational tests are compared to evaluate how costs change in relation to one another. Since the problem is analysed by a mixed-integer linear 57 programming formulation, these experiments also aim at identifying the possibility of solving such problems to optimality using off-the-shelf optimisation software. 06/09/2012 : 14:00 : JM Duddingston Code: OR54A738 Climate Change – A Challenge for MCDA? Development and Use of the UNEP MCA4climate Framework Prof Valerie Belton (University of Strathclyde) The UNEP website describes climate change as one of the major challenges of our time, giving rise to impacts that are global in scope and unprecedented in scale. It advocates anticipatory action today as future adaptation will become more difficult and costly. The UNEP sponsored MCA4climate initiative provides a framework to assist governments in preparing climate change mitigation and adaptation strategies. At the heart of the framework is a generic MCDA model which provides the basis for the evaluation and prioritisation of potential policy actions. The paper will describe the process of model development, which drew on in depth expertise in 12 mitigation and adaptation themes and incorporated three illustrative case studies, and discuss the challenges to MCDA posed by this ambitious initiative. A collaborative project, led by the INCAE Business School in Costa Rica, is currently underway to use the MCA4climate framework to explore the adaptation of the agricultural sector in Peru; progress, challenges faced and hopefully successes achieved will be reported at the conference. 58 Health Organiser: Sonya Crowe 04/09/2012 : 11:00 : JM Prestonfield Code: OR54A531 Demonstrating the Benefit of Operational Research in Maximising the Use of Existing Health Related Data for the Welsh Government Dr Tracey England (Cardiff University), Prof Paul Harper (Cardiff University) and Ms Sarah Lowe (Welsh Government) In October 2011, the Welsh Government recruited four Knowledge Transfer Partnerships Research Fellows (from Cardiff University (2), Swansea University, and London School of Economics) to investigate how existing administrative and survey data could be used more effectively (the fellowships were jointly funded with the ESRC). The role of the Operational Research fellow was to identify areas within the Welsh Government where O.R. techniques (especially computer modelling and simulation) could make use of existing data, highlight data that would be useful in providing policy with evidence, or develop models in a relatively data-free context in order to aid decision making. The overall aim of the fellowship was to deliver projects to demonstrate the added benefit of using simulation or computer modelling with existing Welsh Government data or knowledge. Progress with two specific case studies will be discussed. Case Study 1: Using O.R. to assess the benefits of using teledermatology across Wales. Teledermatology is the use of technology (e.g. videoconferencing, ‚store and forward‛ image transfer) in the diagnosis and triage of dermatology patients. The project considers the benefit of using teledermatology from both the patients’ and clinicians’ perspective. A simulation model is developed to represent the clinicians working schedule. A cost-benefit analysis is undertaken to assess the benefit of reduced patient travel distances and costs. Case Study 2: Using O.R. to assess the benefits of introducing a ‚Direct Access to a Midwife‛ approach to maternity care throughout Wales. ‚Direct Access to a Midwife‛ is an approach where a pregnant woman goes direct to her midwife (on realising she’s pregnant) rather than going to the GP or other health professional. The project considers the initial pathway of the woman and the effect that ‚Direct Access to a Midwife‛ has on her subsequent pregnancy path and associated outcomes. 04/09/2012 : 11:30 : JM Prestonfield Code: OR54A646 Balancing Intensive Care Bed and Theatre Capacity for Elective Cardiothoracic Surgery Prof John Bowers (University of Stirling) Cardiothoracic surgery demands many expensive resources but notably operating theatre time and stay on an Intensive Care Unit (ICU). This study examined the balance between these resources in a specialist facility providing elective heart and lung surgery. Although the admissions can be managed, there are significant stochastic components, notably the cancellation of theatre procedures and patients’ length of stay on the ICU. A simulation was developed to understand the trade-offs between theatre and bed utilisation. This helped staff explore options for expanding the overall capacity for cardiothoracic surgery. Ideally the bed and theatre capacity should be well balanced but unmatched increases in the capacity of either resource can still be beneficial. The study provides an example of a capacity planning problem in which there is uncertainty in the demand and availabilities of two symbiotic resources. 59 04/09/2012 : 12:00 : JM Prestonfield Code: OR54A688 Designing Quota Size and Patient Arrival Time in Multiple Block Systems with Limited Medical Resource Dr Kwei-Long Huang (National Taiwan University) and Miss Jei-Fen Tsai (National Taiwan University) Two performance measurements are considered for patients who wait for the required service to be fulfilled: one is the period started from the date of requesting the service to the appointment date, and the other is the time from entering the facility to getting the service started. For critical and scarce resources in healthcare, limited quota for these services are provided in each appointment time slot. To utilise the resources and enhance the two measurements of waiting time, we propose a 2-stage appointment policy to satisfy patients’ appointment requests and preferences. The proposed policy provides a patient with information about which date and at what time he should arrive at the facility. The proposed policy takes patients’ characteristics such as the degree of urgency and preferences of time slots into consideration as well as the service times of different types of patients. Stage 1 of the proposed appointment policy is to address a daily-base appointment scheduling which assigns a patient to a time slot in a date. In this stage, a score of ‚Balance Utilisation‛ of the system and ‚Patient Preference‛ for each available time slot is calculated and the slot with the highest score is assigned to the requesting patient. Furthermore, in order to determine the arrival time of a patient entering the facility, a time slot is divided into multiple time blocks with shorter time. Given a limited waiting time for patients, a probability that patients may wait more than that tolerance time is calculated so that a quota of serving patients in each time block is determined by dynamic programming. In Stage 2, based on the quota size of each block, the arrival time of a patient is confirmed. The proposed 2-stage appointment policy is evaluated by using simulation on a real case. 04/09/2012 : 13:30 : JM Prestonfield Code: OR54A656 Applying Temporal and Social Cues to Information Design in Health Communications Miss Sabrina Pei-Yi Cheng (National Taiwan University) and Dr Jiun-Yu Yu (National Tawain University) Reaching an all-time high of over 10% of GPD of most developed nations in 2011, health expenditure is expected to continue to grow. On one side, as technology advances, health service providers focus more strongly on various segments and develop diversified products and services to satisfy different types of consumer demand. On the other side, health consumers deal with multiple appeals, from disease control, illness prevention, to health promotion, and try to find out what the information reveals. Generally, in such communication embedded in the health services delivery process, there would be a gap between service providers and consumers, so information design consequently plays an important role in this progressing industry. Health service providers usually employ statistical data to convince message receivers of the benefits and effects or to give weight to the health issue. Such information presentation often acts as temporal and social cues for creating a linkage between personal experience and outcome of adoption the health products, services, or even behaviour suggestion. For example, how many deaths in a period, when the effect begins, who may be affected, etc. Therefore, the current research tries to seek out the appropriate message layouts not only based on health industrial practices but also ensuring that the information is realisable and useful for receivers. This research compromises time ” number of deaths per year/per day, and healthy food intake for weeks/months ” and role ” self, family, and friends ” via collection and analysis of the empirical data. The results confirm a new effective strategy in health operations communication by showing the positive effect of information congruency on receiver evaluations. Thus, health service providers can distinguish the more powerful message from commonplace one, and information receiver can handle easily and fluently while face increasing information volume in health industry. 60 04/09/2012 : 14:00 : JM Prestonfield Code: OR54A705 Phasing the Introduction of a Revised General Practitioner Funding Formula Mr Orville D'Silva (), Dr David Worthington (Lancaster University) and Dr James Crosbie (Department of Health) General practices under the General Medical Services (GMS) contract currently receive substantial payments from the Department of Health based on the Carr-Hill formula. A review of this formula was carried out by the Formula Review Group (FRG) in 2005 who concluded that a revised formula, known as the FRG formula, was needed. However, the FRG formula was not implemented at the time. Both formulae are made up of six indices which are designed to reflect practices’ workloads and the environments they operate in, and therefore costs within general practice. Simple analysis showed that an instantaneous switch from the CarrHill formula to the FRG formula would mean that some practices would face single-step changes in funding which would clearly represent significant and problematic change in incomes. Thus, the Department of Health is exploring a transition period during which the components of the Carr-Hill formula are gradually replaced by the components of the FRG formula. This paper describes a project undertaken for the Department of Health to develop and use heuristic algorithms to obtain the ‘optimal’ order in which the components of the two formulas may be phased in and out. Lessons learned from the case study about the use of heuristics in practical situations are also drawn. 04/09/2012 : 15:30 : JM Prestonfield Code: OR54A652 A System Dynamics ‚Flight Simulator‛ for the Evaluation of Policy Interventions in Patient Pathways for Cutaneous Malignant Melanoma Mr Paul Holmstrom (Chalmers University of Technology), Mrs Magdalena Claesson (Sahlgrenska Universitetssjukhuset) and Mr Stefan Hallberg () A system dynamics simulation model was developed as part of building improved long-time planning and policy evaluation in the patient pathways for Cutaneous Malignant Melanoma. An ‚incidence generator‛ was used to create patient output sets dependent on different patient delays in seeking treatment. A ‚flight simulator‛ was built to test the scenarios against incidence rates, follow-up programmes etc, assessing the trajectories of healthcare costs over time. It is shown that costs escalate significantly unless patient delays are addressed and follow-up programmes reviewed. 04/09/2012 : 16:00 : JM Prestonfield Code: OR54A730 Mechanism Design for Examination Resource Allocation in a Healthcare Organisation Dr I-Hsuan Hong (National Taiwan University), Dr Wen-Chih Chen (National Chiao Tung University) and Mr Kun-Liang Lien (National Taiwan University) The healthcare industry has drawn much attention recently. The examination resources are expensive and limited. Hence, the issues of resource allocation and use in healthcare organisations are interesting and important. Typically one dedicated department is established for examination among different departments to integrate examination resources. The purpose of this research is to design a resource allocation mechanism in the perspective of a healthcare organisation. We investigate the incentive mechanism design problem for resource allocation in a healthcare organisation so that the individual behaviour of departments can coincide with the goal of the healthcare organisation. The Vickrey-Clarke-Groves (VCG) mechanism is widely applied to auction practices and the true willingness to pay of the bidders can be revealed under the VCG mechanism. We apply the VCG mechanism on the resource allocation problem in a healthcare organisation so that departments are willing to report the true need of the examination resources by the mechanism of paying an internal transfer price. Then, we conduct a sensitivity analysis to examine how each type of the examination resources influences the total payoff of a healthcare organisation. Finally we show a case study to demonstrate the proposed mechanism. 61 04/09/2012 : 16:30 : JM Prestonfield Code: OR54A655 Process Improvement and Performance Measurement for Emergency Medical Services (EMS) Dr Jiun-Yu Yu (National Taiwan University) Emergency Medical Services (EMS) refers to both the patient transport and the medical support solution for people with urgent illnesses or injuries. New Taipei City (NTPC) is the most populated region in Taiwan, and the demand for EMS in NTPC is expected to grow at a speeding rate. Thus it is vital to upgrade the current EMS system of NTPC to an integrated and well-managed state. Most past research about EMS is conducted from the clinical perspective, focusing on the benefits that EMS brings to the patients. Believing that the patient’s health benefits can be enhanced operationally with the help of a better designed and implemented EMS process, the author of this study proposes a framework based on an integrated approach. Current EMS at New Taipei City practices are empirically observed and interviewed, the results of which are further investigated by process analysis and statistical data analysis. Service operations management tools, such as service blueprint and service profit chain, and the model of balanced scorecard (BSC) are applied to guide the construction of the key performance indicators (KPIs). System thinking is then employed to explore the causal structures among the KPIs, aiming to identify achievable potentials for process improvement and performance measurement. 05/09/2012 : 08:30 : JM Prestonfield Code: OR54A606 KEYNOTE: O.R. and the Industrialisation of Healthcare Prof Michael Pidd (Lancaster University) As should be obvious from previous OR Society conferences, O.R. has a long and successful history in healthcare, covering a range of application areas. Recent years have seen the increasing popularity of approaches to healthcare organisation, of which lean is the latest, that began life in manufacturing industry. This paper attempts a brief survey of O.R. applications and developments in healthcare, and considers where O.R. analysts can add the greatest value. It takes a realistic look at the probability of successful applications and a critical look at some academic developments. It considers the degree to which O.R. is part of a movement to standardise and industrialise healthcare, and ponders the value of this against a scenario in which much healthcare will no longer be delivered in hospitals. 05/09/2012 : 11:00 : JM Prestonfield Code: OR54A610 Putting Research into Practice: A Pilot Project to Implement a New Way for UK Paediatric Cardiac Units to Monitor Their Short-term Surgical Outcomes Dr Christina Pagel (University College London), Dr Kate Brown (Great Ormond Street Hospital), Dr Sonya Crowe (University College London) and Prof Martin Utley (University College London) Adult cardiac units in the UK routinely monitor their survival outcomes adjusting for the severity of their case mix, supporting improvements in the services they offer. However to date there has been greater difficulty in evaluating the (short-term) mortality outcomes in paediatric cardiac surgery. While the UK Central Cardiac Audit Database (CCAD) currently publishes outcomes by procedure type for all units, outcomes across a whole programme are not presented because of the difficulty of adjusting for the severity of case mix in children. We have just completed a pilot project to implement such programme-level monitoring in UK paediatric cardiac units for the first time. In collaboration with Great Ormond Street Hospital and the CCAD we developed a new risk model for 30 day mortality following paediatric cardiac surgery, finishing in October 2011. From the beginning of this project, we were clear that we wanted the risk model to be used by paediatric cardiac units in-house to monitor their own outcomes in a routine manner, mirroring current adult practice. Thus, throughout the project we engaged the clinical community, presenting updates on progress at national meetings. The interest this generated meant that almost as soon as the model was developed, we had UK centres eager to pilot the model for routine risk adjustment in house. We started the pilot in November 2011 and worked closely with pilot units to develop Excel software that could be used by data managers to implement our risk model and produce charts showing the unit's risk-adjusted 30-day outcomes over time. The pilot project has now finished and pilot units are continuing 62 to use the developed prototype software. Key to the success of this pilot was taking the time to engage with both data managers and clinicians at each unit. 05/09/2012 : 11:30 : JM Prestonfield Code: OR54A666 Sequential Drug Decision Problems in Long-term Medical Conditions: A Systematic Review on Potential Heuristic Methods Miss Eun-Ju Kim (School of Health and Related Research, University of Sheffield), Mrs Roberta Ara (School of Health and Related Research, University of Sheffield), Mr Jim Chilcott (School of Health and Related Research, University of Sheffield) and Dr Yang Meng (School of Health and Related Research, University of Sheffield) Sequential drug decision problems (SDDP), which assign optimal drug(s) sequentially to the disease pathway over time, are important not just for clinical decision-making but also for efficiently managing the constrained resources for long-term medical conditions such as hypertension. However, solving a real world SDDP can be complex because of: the large number of alternative treatment sequences; the large number of combinations of multiple health states; the interaction between the treatment sequences and the health states and the uncertainty surrounding the problem. We conducted a systematic review to provide a comprehensive insight into the potential heuristic or optimisation methods for identifying optimal or nearoptimal treatment sequences in a large decision space. Four databases (Web of Science, Scopus, Medline and Embase) were searched for studies published between January 1990 and March 2012 that addressed a sequential optimisation problem using either a heuristic or optimisation method. Of the 10,517 initially retrieved studies, 543 studies, which satisfied the predetermined inclusion and exclusion criteria, were classified according to the method used: mathematical programming, constructive methods, local search methods or other heuristic methods. The initial findings suggest that dynamic programming and some local search methods (e.g. genetic algorithm, simulated annealing) are potential promising methods to deal with SDDPs. The performance of these two types of methods may depend on whether an exact optimal solution of an approximate model (dynamic programming) or an approximate near-optimal solution of a more detailed model (local search methods) is preferred. In further work, several local search methods identified in the systematic review will be tested using the case of antihypertensive treatment. 05/09/2012 : 12:00 : JM Prestonfield Code: OR54A695 SoApt: Service Option Assessor and Prioritisation Tool Mrs Carol Marshall (University of Stirling) The Service Option Assessor and Prioritisation Tool is based on the principles of Multi-criteria Decision Analysis (MCDA) and Programme Budgeting and Marginal Analysis (PBMA) to assess and prioritise investment proposals for health and social care. SoApt, utilising the six domains of quality as criteria, calculates the impact on the care of the user and the impact on the costs to the organisation of a new or redesigned project. The criteria are weighted to reflect local and national priorities. The proposals are captured and compared on a matrix which allows the decision-maker to identify which proposal(s) to invest (or disinvest) in. A successful proposal generates a Status Report within the model. The Status Report monitors progress against the predetermined quality and cost objectives and is completed by the proposer and assessed by the decision-maker. 06/09/2012 : 08:30 : JM Prestonfield Code: OR54A651 Simulation for Sustainability in Healthcare: Identifying Gaps in the Literature Mr Masoud Fakhimi (Swansea University), Dr Navonil Mustafee (Swansea University) and Dr Jane Probert (Swansea University) Factors relating to sustainability are becoming an increasingly integral part of managerial decision making processes; there is a growing awareness among the stakeholders that their success is profoundly dependent on creating a harmony between triple bottom line (an integration of economic, social and environmental responsibilities) with respect to their strategic priorities. Thus, it comes as no surprise that the scope of 63 systems’ modelling that hitherto focussed predominantly on productivity-related performance measures is gradually expanding to incorporate sustainability-related criterion. Existing publications indicate that there has been a dearth of empirical research on integrating sustainability factors with modelling; this is especially true in healthcare. Thus, this research has employed a cross-industry review of literature on modelling for sustainability with the objective of identifying studies that could be applied in the healthcare context and in identifying gaps that may exist. The review of literature has shown an unequal treatment of economic, social and the environmental factors among studies employing qualitative models (e.g. conceptual models) and those using quantitative/mathematical modelling (e.g. computer simulation). Whereas the former modelling approach has considered the three aforementioned sustainability-related factors in the formulation of guidelines, frameworks, best practices, etc., the latter has mostly ignored the societal aspects of sustainability and has principally focussed on the economy and the environment. This research argues that the level and range of sustainable considerations in model development are different in healthcare and non-healthcare industries. Healthcare is a welfare-oriented service for citizens so, although quantitative modelling studies in non-healthcare domains such as manufacturing have frequently neglected the social factors pertaining to sustainability, healthcare modelling cannot ignore this essential aspect. Informed by the literature review, the research has also identified some challenges in sustainable healthcare modelling, and investigates commonalities and differences between healthcare and nonhealthcare modelling for sustainable success. 06/09/2012 : 09:00 : JM Prestonfield Code: OR54A662 Supporting Pre-planning Design Phases of New Dementia Care Environments through Group-modelling Mr Paul Holmstrom (Chalmers University of Technology), Dr Marie Elf (Dalarna University College), Dr Inga Malmqvist (Chalmers University of Technology), Prof Kerstin Öhrn (Dalarna University College) and Dr Lena von Koch (Karolinska Institutet) The pre-design phase, before architectural design details are presented, is important for the success of the entire construction process. In this phase, it is important to identify the needs of the users and the organisation. The present project describes a group-modelling project with system dynamics (SD) in the predesign phase of a new dementia care environment. The aim was to support a dynamic design process that was oriented towards the users’ needs. SD uses models to create scenarios and experiment with different solutions to understand complex problems. The method is interactive and suited for group- modelling. The tools used are a construction of a diagram depicting the system and a mathematical model for scenario experiments. The group- modelling was performed during five workshops with stakeholders from the organisation. The data for the model was collected through discussions with the modelling group, a questionnaire, literature reviews and documents. The process was documented by video-recording and field notes. The participants were interviewed at the end of the project. The data was analysed by content analysis. Results show that the modelling process supported the group to explicitly express their goal regarding care. The modelling process moved the group discussions from space solutions and square meters to the goals of the care process. The model developed illustrates a modern perspective of dementia care. Group-modelling using SD is interesting for the pre-design process. It can be feasible for linking strategic planning of the organisation with the plan over the new healthcare environment since it moves the participants in the pre-design phase to focus on important issues about the care and away from a wish list of various needed spaces, commonly achieved in traditional planning processes. 06/09/2012 : 09:30 : JM Prestonfield Code: OR54A673 Are We Over-analysing Health Care and Health Services? Ms Penelope Mullen Regrets, even criticisms, have long been expressed that much analysis within health care and healthcare systems, both in O.R. and in related areas such as health economics, has focused on ‘micro’ applications ” single processes, new technologies/therapies, specific hospitals, individual departments within hospitals etc. It has been claimed that few studies focus on the whole or wider healthcare system or have wide-scale 64 application - leaving some with a feeling of ‘chipping away at the edges’ whilst ignoring the essential core. There have, of course, been some welcome developments. Recent years have seen the long overdue pursuit of generic models, which can be adapted and applied in different settings. Department of Health analysts have contributed to a range of system-wide developments, such as NHS Direct. Possibly more controversially, maximising health gain as the sole, or at least dominant, health-care system objective, championed by health economists, has found advocates across the (English?) NHS. However, do systemwide analyses, whether large-scale or small-scale with universal coverage, risk over-analysis ” the imposition of apparently rational and systematic approaches on the interdependent, ever-changing, professionally-led, messy combination of activities and services which constitute health care. Would such analyses alleviate or would they contribute to fostering the fragmentation of services into discrete elements or chunks occurring, at least in part, in response to the demands of competitive commissioning? Could analysis to support system-wide investment/disinvestment decisions undermine the objectives of a universal health service? Is there a danger of distortion resulting from large-scale focusing on the ‘easy pickings’ leaving the ‘messes’, and their associated services and patients, behind? What are the implications of ‘system-wide’ analysis in the presence of a goal of maximising health gain? This paper asks whether largescale or system-wide application of ‘rational’ analysis would mitigate or exacerbate such problems and, possibly, presents some answers. 06/09/2012 : 11:30 : JM Prestonfield Code: OR54A726 Modelling the Dental Workforce in Sri Lanka Prof Sally Brailsford (University of Southampton) and Dr Dileep De Silva (University of Southampton) This talk presents a system dynamics simulation model of the supply and demand for dental care in Sri Lanka. The model represents the flows of dentists through the complex recruitment and career progression phases through to retirement. This is augmented by a demand model, which combines empirical data on the time required to carry out various dental procedures with secondary clinical and demographic data to project the future incidence of different dental conditions. The overall supply-demand model enables policymakers to test various policies for staffing, training and recruitment. 06/09/2012 : 12:00 : JM Prestonfield Code: OR54A702 Using Data Mining and Simulation for Health System Understanding and Capacity Planning: An Application to Urgent Care Mr Mohamed Tadjer (Hounslow Primary Care Trust) Many visits to Accident and Emergency (A&E) are made for problems that do not require emergency interventions. Care for non-urgent cases is more costly in the A&E than in primary care settings. Various methods have been used in attempts to decrease non-urgent A&E use. Often, triaging method is used to redirect patient with non-urgent care problems to primary care facilities. This method requires health professionals to assess the urgency of the patient's problem before determining whether Emergency care is appropriate. The purpose of this study is to assist Hounslow Primary Care Trust examine their Accident and Emergency services and to help improve health access by redesigning these services into two different settings: Urgent Care Centre (UCC) and Emergency Department (ED). This study used data mining techniques to understand system behaviour in terms of subsystem interactions and the factors influencing patient flows. The insight obtained through clustering analysis is then used as patients’ demand and pathways supporting better services in the UCC. The data have been clustered into groups and sub groups on their resources used and acuity. These groups are the identified target groups for service specification strategy design and decision making. We used simulation modelling to help understanding the correlations that exist between various inputs into a UCC (i.e., capacity and resources, facility and staff resources) and various output performance measures from the system (i.e., patient waiting times, staff utilisation, resources and facility utilisation). 65 06/09/2012 : 13:30 : JM Prestonfield) Code: OR54A566 MCDA Approaches to Prioritisation in NICE Mr Brian Reddy (University of Sheffield) The National Institute for Health and Clinical Excellence (NICE) is an agency of the British National Health Service (NHS), providing guidance to the health services and other relevant decision making bodies in England regarding new drugs and technologies, broader clinical practice and public health. It attempts to reduce variations in levels of treatment between local regions and encourage best practice across the health system with evidence based approaches to decision making. However, for a number of reasons, it is challenging to adequately quantify, model and describe the multiple effects of public health interventions. As a result there is some inevitable uncertainty around their cost effectiveness, which may lead to inappropriate decision making regarding which interventions to recommend. This presentation will describe some of the most common difficulties of measuring cost effectiveness in public health and explain why multi-criteria decision analysis (MCDA) techniques may be appropriate to overcome some of these issues. It will then explain an example of its use within the Centre for Public Health Excellence (CPHE). An Analytic Hierarchy Process approach was used to prioritise which topics should be selected for future guidance. Further examples of expected approaches to be trialled or incorporated will then be discussed. 06/09/2012 : 14:00 : JM Prestonfield Code: OR54A593 Optimal Location and Scheduling of Blood Collection Facilities Mr Thomas Jeffries (University of Kent) and Dr Jesse O'Hanley (University of Kent) The collection, processing and distribution of blood products are vital parts of any healthcare network. Although much research has investigated the processing and distribution stages, there has been little research focusing on the collection stage, specifically, how a blood service should organise the collection of blood donations from the general population. As blood has a relatively high unit cost and a limited shelf life, as well as the fact that blood services have finite resources, it is important that a blood service does not over or under collect blood, resulting in excess wastage or, worse yet, putting patients at risk. Taking such factors into account, it is vital that a blood service operates a highly efficient blood collection schedule. This talk considers how a blood service should efficiently organise their blood collection schedule over a multiperiod planning horizon. We develop an optimisation model to determine: how many, how frequently, of what type and where donation sessions should be located in order to minimise overall cost. Importantly, the model takes into account resource constraints on the blood service, the predicted demand for each blood group, how the different blood groups are distributed throughout the population and required male/female donation frequencies. Results of our analysis for the South East of England demonstrate the efficiency of the proposed model over current blood collection scheduling procedures. 66 Horizon Scanning & Future Analysis Organiser: Amanda Tucker 04/09/2012 : 13:30 : JM Pentland East Code: OR54A696 The Ten Commandments of Horizon Scanning Mr John Carney (Dstl) The UK Government definition of horizon scanning (2004) is 'the systematic examination of potential threats, opportunities and likely developments including but not restricted to those at the margins of current thinking and planning. Horizon scanning may explore novel and unexpected issues as well as persistent problems or trends'. All UK Government departments have discharged their remit for an horizon scanning function in different ways and within Dstl (the Ministry of Defence’s in-house science and technology laboratory) a bespoke process has been designed aimed at identifying potentially disruptive, and relatively unknown areas of science and technology, with potential relevance to defence and security. As the practice of horizon scanning has become established across Government so the expectation for exploitation and demonstration of impact by stakeholders has increased. This paper will recount some of the perceived pitfalls around exploitation and common misunderstandings that can occur, and will cite specific examples of the challenges therein. This paper represents the personal perspective of the author as a member (and erstwhile Team Leader) of Dstl’s Horizon Scanning Team, combined with the perspectives from colleagues in Other Government Departments, whose contribution is gratefully acknowledged. 04/09/2012 : 14:00 : JM Pentland East Code: OR54A745 Multiple Crystal Balls – Preparing for an unpredictable future Miss Hannah Locke (Dstl) In defence, we are concerned with ensuring vital capabilities will be met in the future. But how do we know what capabilities will be required in order to prepare for them if we don’t know what our future world will look like? Will access to resources continue to decline or will innovation overcome our current issues? Will the Euro recover or will the EU be the first of many international communities to fail? Will Twitter become the corporate communication norm or will increased cyber security put an end to the social networking age? Instead of trying to predict the future, the Conceptual Force Development programme, run by the UK Ministry of Defence, aims to overcome the issue of uncertainty by looking at a variety of different possible futures for 2030 and beyond. By exploring the numerous key themes, both worldwide and UK specific, for each future under observation, and identifying which issues and insights are specific to one future and which will be likely in a range of futures, the study has begun to highlight potential capability gaps within the UK defence force which we can then help to inform and prepare for. 67 04/09/2012 : 14:30 : JM Pentland East Code: OR54A748 Systems Thinking, Scanning, and Scenarios: Exploring Future Landscapes by Identifying Critical Interconnections. Dr Wendy L Schultz (Infinite Futures) How can systems thinking, including influence and causal loop mapping, contribute to research techniques in futures studies and foresight? A focus on interconnections among trends and emerging issues, and resulting impact cascades, can identify 'missing changes' - spaces where change has not yet emerged, but which create the heightened potential for change. Systems thinking also provides a conceptual framework for mapping impact cascades resulting from emerging change, as seen in foresight tools such as implications wheels. Combining implications wheels with causal loop mapping enables a robust new approach to inductive scenario building. This presentation will offer both an overview of these methods and their connections to systems thinking, and case studies and examples of their use in both government and business sector organisations. 68 Information Systems & Knowledge Management Organiser: Jo Smedley 04/09/2012 : 11:00 : ST.L.Bonnar Code: OR54A553 Monitoring the Service Operations of a Large Financial Enterprise using a Systems Integrator Approach to Alerting and Event Management: A Case Study of a Large Australian Bank Mrs Caroline Sanders (Westpac Banking Corporation) The paper will outline the overall ITIL Service Management strategy that was developed but will focus on the enterprise monitoring and alerting capabilities that form the foundation of good event management. A target state will be described and the integration approach taken to ensure a successful widespread roll out of monitoring agents across a large number of configuration items (both infrastructure and application) that comprises the system landscape. Outcomes both in terms of improved processes and key performance indicators, such as mean time between failure and mean time to restore service will also be described. The use of these capabilities by both internal and outsourced service providers and by a centralised National Operations Room will be considered. 04/09/2012 : 13:30 : ST.L.Brewster Code: OR54A561 Weaving the Innovation Web: Using Knowledge Management and Information Systems to Develop Communities of Practice Dr Jo Smedley (University of Wales, Newport) Knowledge management supported by information systems help to structure and communicate complex problems to gain a common understanding among a group and progress on a committed course of action. This paper illustrates how a combination of soft and hard O.R. approaches supported the design, development and implementation of an organisational learning and teaching initiative across a UK University. Hard O.R. is usually regarded as tangible, easy to explain and easy to use. In comparison, soft O.R. is somewhat intangible, not easy to explain and not easy to use. O.R. as a collaborative learning "language" across disciplines and topics supported effective and efficient engagement helping to achieve successful outcomes and inform continuing developments. 04/09/2012 : 14:00 : ST.L.Brewster Code: OR54A568 The Secret to Preparing your Data for Advanced Analytics Ms Sayara Beg () Rubbish data in, rubbish data out. If the data is not properly organised in the first place, the results of your advanced analysis will be meaningless, however, you cannot start by trying to 'boil the ocean', so where should you begin? Many organisations think that reference data is the place to start and indeed have spent large amounts of much needed cash on reference data related projects to support data warehouse initiatives or business intelligence programmes with very little return on that investment, are you one of them? Did you know there is another way, a much more cost effective way, to organise your data? A way that doesn’t require cost-prohibitive, labour-intensive and time-consuming reference data projects to be initiated? What is this secret? I will show you, how, through a process of scoring, weighting and categorisation of your raw data you will not only meet your expectations but you will receive a richer, 69 deeper understanding of your own business and your customer’s behaviour using technology-agnostic advanced analytical methods and you will not need to tidy up your data. Well you feel the urge to do some tidying up, simply because you have been told for so long that you need to. However after I've revealed the secret to organising your data in preparation for Advanced Analytics, you may decide not to bother trying after all. 04/09/2012 : 14:30 : ST.L.Brewster Code: OR54A712 Knowledge Management and Sustainable Quality Dr Brian Lehaney (Quality Assurance Agency for Higher Education) Sustainable quality is relevant to organisations and individuals working in any sector of the economy. Knowledge management has become a key process in understanding organisations and their use of resources and, ultimately, quality is a major differentiating factor when considering goods and services. Sustaining quality requires taking a strategic view that may present short to medium term challenges and knowledge management should be able to help address such challenges. For large organisations, knowledge management may be seen as an intra-organisation activity, but sustaining quality for small to medium enterprises may require inter-organisational cooperation. The presentation addresses quality and knowledge management issues faced by different sectors organisations, and how these are addressed in practice and in theory. 70 Logistics and Supply Chain Organiser: Patrick Beullens 04/09/2012 : 15:30 : ST.L. St.Trinneans Code: OR54A560 Assessing End User Acceptance Risk for Military Systems Prof Patrick Driscoll (U.S. Military Academy), Dr Timothy Elkins (U.S. Military Academy) and Mr James Enos (U.S. Military Academy) End user rejections of well-engineered technological systems coupled with perceptions of 'drive-by fielding' confound rapid acquisition support for deployable force protection for U.S. ground forces. Linking preacquisition assessments directly to end user expectations within a tradespace defined by technological, user, and logistics factors, offers significant insights that are informing system design choices and acquisition risk management decisions in a way that offers potential solutions to both these problems. 04/09/2012 : 16:00 : ST.L. St.Trinneans Code: OR54A679 Supply Chain Competition with Capacity Constraints Dr Shu-Jung Sunny Yang (University of Essex), Prof Eddie Anderson (University of Sydney) and Dr Yong Bao () Capacity constraints most often determine the nature of competition between companies and have a major impact on their profitability. In this paper, we examine the role of capacity information in a supply chain with one supplier and two downstream sellers. We find that in the absence of transferable capacities between both sellers, for most cases it is better for the supplier to release its capacity information so the sellers can make better ordering decisions. This outcome holds under the existence of transferable capacities as well. We also demonstrate in our analytical model that the value of capacity information is zero when there is ample capacity. These results may help to explain why capacity information is always announced prior to the ordering decision in capacity constrained supply chains such as in semiconductor manufacturing. 04/09/2012 : 16:30 : ST.L. St.Trinneans Code: OR54A690 The Orienteering Problem with Variable Profits Dr Gunes Erdogan (University of Southampton) and Prof Gilbert Laporte (HEC Montreal) This paper introduces, models and solves a generalisation of the Orienteering Problem, called the The Orienteering Problem with Variable Profits (OPVP). The OPVP is defined on a complete undirected graph G = (V,E), with a depot at vertex 0. Every vertex i ∈ V \ {0} has a profit p_i to be collected, and an associated collection parameter α_i ∈ [0, 1]. The vehicle may make a number of ‚passes‛, collecting 100α_i percent of the remaining profit at each pass. In an alternative model, the vehicle may spend a continuous amount of time at every vertex, collecting a percentage of the profit given by a function of the time spent. The objective is to determine a maximal profit tour for the vehicle, starting and ending at the depot, and not exceeding a travel time limit. A linear integer programming formulation for the former model and a nonlinear integer programming formulation for the latter are provided. Linearization schemes for the nonlinear models for the cases of concave and convex collection functions are devised. Results show that the discrete pass model can be solved for about 200 vertices within two hours of computing time, whereas the continuous time model is beyond the computational reach for more than 75 vertices. 71 05/09/2012 : 08:30 : ST.L. St.Trinneans Code: OR54A534 Quality Assessment, Procurement, and Remanufacturing Decisions in a Reverse Supply Chain with Multiple Returns Collection Locations Mr Christos Zikopoulos (Aristotle University of Thessaloniki) and Prof George Tagaras (Aristotle University of Thessaloniki) Exploitation of the value remaining in used products after their end-of-use has been proven in many cases as a profitable operation. As a consequence of the associated economic benefits and the interrelated environmental legislation, currently there is a diversity of products that flow through the reverse channel of modern supply chains in order to undergo some kind of recovery action, i.e. reuse, repair, remanufacture. The management of reverse supply chains shares some common features with the conventional (forward) supply chain management. However, there are certain characteristics which, although unimportant or even absent in the conventional supply chain context, are extremely significant when it comes to reverse supply chain management. Among these diversifying features one can include the increased product quality variability and uncertainty, as well as the availability of a multiplicity of locations where used products are collected in order to supply the production system. Uncertain quality of returns, apart from increasing the complexity of procurement and recovery decisions, may dictate the establishment of procedures for the quality assessment and classification of returns. The importance of explicitly dealing with quality variability and the necessity of returns grading and sorting are evident in both recent academic research and current business practices. In the current paper, we examine the problem of procurement and remanufacturing lot sizing in remanufacturing systems supplied from multiple returned products collection facilities in a single period context under alternative returns classification configurations. Our scope is to examine the impact of the availability of multiple collection locations on reverse supply chains profitability under different assumption regarding the characteristics of returns quality at these locations. Moreover, we study the conditions that determine the most advisable classification procedure among a number of alternatives that are used in some real-life reverse supply chains. 05/09/2012 : 11:30 : ST.L. St.Trinneans Code: OR54A602 A Technique to Develop Simplified and Linearised Models of Complex Dynamic Supply Chain Systems Ms Virginia Luisa Marques Spiegler (Cardiff Business School), Prof Mohamed M. Naim (Cardiff Business School), Prof Denis R. Towill (Cardiff Business School) and Prof Joakim Wikner (Link•oping University) Non-linearities are features of most real-world dynamic systems. The existence of non-linearities makes the system analysis and synthesis very difficult. For this reason, previous work in supply chain dynamics literature has mainly focused on examining ‘presumably linear’ models. In addition to this, averaging techniques and disregarding discontinuities have been common practice for approximating non-linear systems. However, such linear assumptions and approximations do not always yield the necessary degree of accuracy required. This work aims to identify and categorise different types of non-linearities that commonly appear in supply chain dynamics models, as well as to compare and contrast suitable methods for analysing each type of non-linearity. We also suggest some simplification methods to reduce model complexity and to assist in gaining system dynamics insights. Hence, the purpose of this paper is to develop more accurate simplified linear representations of complex non-linear supply chain models. We use the well-known Forrester production-distribution model as a benchmark supply chain system to study non-linear control structure and apply linearisation methods. Within the control engineering literature such methods include small perturbation theory, describing functions and phase plane analysis. We then compare performances of the linearised models with numerical solutions of the original Forrester model and with previous work. Initial findings suggest that more accurate linear approximations can be found for reproducing the original Forrester model. These simplified and linearised models enhance the understanding of the system stability, reliability and actual transient responses between the steady-state operating points. We also find that the use of simulation and numerical techniques complements the study of non-linearities and helps to illustrate the sensitivity behaviour of a supply chain system under uncertain operating points. A step by step 72 procedure is provided for the analysis and design of non-linear supply chain dynamics models, especially when overly simplistic linear relationship assumptions are not possible or appropriate. 05/09/2012 : 12:00 : ST.L. St.Trinneans Code: OR54A691 The Capital Costs and Rewards of Inventories in the Supply Chain: A Generalised MultiEchelon Inventory Theory Dr Patrick Beullens (University of Southampton), Prof Gerrit K. Janssens (University of Hasselt) and Prof Luk N. Van Wassenhove (INSEAD) Classic deterministic multi-echelon inventory theory is based on an echelon-specific unit holding cost, and is used to find optimal lot-size decisions that minimise the total inventory holding costs in a supply chain. In this paper, we generalise this multi-echelon inventory theory to the case that each echelon represents an independent firm, which is an important consideration in the supply chains of today. In this context, it is essential to obtain the impact of lot-size and inventory decisions on the individual profit functions of each of the firms involved. We show how from the set of individual profit functions derived, traditional multiechelon theory follows. Conversely, we show that the traditional theory is not sufficient on itself to arrive at the same information. Furthermore, our approach shows that profit functions and optimal inventory decisions are highly dependent on cash-flow assumptions, which the traditional theory never makes explicit in inventory models. As a result it can also be concluded that firms should consider in general at least two different types of unit holding costs and two types of inventories. The distinction between installation and echelon inventory holding costs, which arises in the classic theory, is unnecessary. Our approach thus offers a way to generalise the traditional multi-echelon theory, making it an appropriate framework in which to study supply chain organisation and inventory coordination across independent firms. 06/09/2012 : 09:00 : ST.L. St.Trinneans Code: OR54A611 Effect of Service Time on Service Costs: Comparison of Hierarchical v/s Non-hierarchical Setup in Service Parts Logistics Mr Mohsin Jat (University of Nottingham), Prof Bart MacCarthy (University of Nottingham) and Dr Luc Muyldermans (University of Nottingham) This work is motivated by large scale IT Service Parts Logistics systems supporting provision of service-parts under different and short service time commitments. Typically in IT hardware support contracts, depending on the consequences of equipment downtime, customers determine different time windows (e.g. 2 hours, 4 hours and 8 hours) within which the requested service-part(s) should reach their sites. We model two distinct setups of service facilities, namely hierarchical setup and non-hierarchical setup, to investigate the impact of varying service time limits on inventory, transportation and facility setup costs in a system that supports time-based differentiated services. Cost functions for approximating inventory, transportation and facility setup costs are derived to develop the analytical model. We consider an (S-1, S) inventory policy to determine inventory costs, and assume Euclidean distances and hexagonal packing of service areas to determine travelling distances and number of facilities, respectively. The analysis highlights that hierarchical setup exploits the opportunity to provide service with higher level of centralisation in response to requests for services in relaxed time. This can result in lower overall inventory level in the system, however, at the expense of higher transportation costs. 06/09/2012 : 13:30 : ST.L. St.Trinneans Code: OR54A607 Modelling Supply Chain Networks Reliability and Resilience Using a Dynamic Bayesian Network Mr Mouhamad Shaker Ali Agha (University of Strathclyde ), Dr Robert Van Der Meer (University of Strathclyde) and Prof Lesley Walls (University of Strathclyde) The resilience and reliability of supply networks following catastrophic events such as the 9/11 attack and the Japanese tsunami respectively has received an increased significance. We investigate how supply networks resilience relates to reliability, robustness and risk, in order to construct models to support 73 assessment of supply networks resilience. The usefulness of Dynamic Bayesian Networks to model uncertainties in events affecting supply and how these uncertainties propagate through supply networks over time will be explored. Using simple examples, we examine how our model can be employed to evaluate the impact of different strategies intended to improve supply networks resiliency. 06/09/2012 : 14:00 : ST.L. St.Trinneans Code: OR54A608 Two Phase Optimization of the Reverse Logistics Network with Multiple Recovery Routes and Quality Inspection Mr Ali Niknejad (Coventry University) and Prof Dobrila Petrovic (Coventry University) Attention to optimisation of reverse logistics networks has grown rapidly in the past few decades. In this paper, an inventory control and production planning model for the integrated reverse logistics network with two alternative recovery routes, including repair and remanufacturing, a disposal route and a traditional forward route is introduced. Both demand and return volumes are assumed to be uncertain and fuzzy trapezoidal numbers are utilised to represent them. Quality of returned products is considered during the returned products’ inspection and repair and remanufacturing quality thresholds are determined to segregate returned products into the respective recovery routes. Furthermore, a two-phase optimisation approach is developed and implemented to determine the network decisions as follows: In phase 1, volume of returned products to be sent to repair, remanufacturing and disposal are determined while, in phase 2, decisions regarding time and volume of component procurement, production, repair and disassembly are made using fuzzy optimisation. 06/09/2012 : 14:30 : ST.L. St.Trinneans Code: OR54A634 The Effects of Judgemental Adjustment in an Inventory System Mrs Inna Kholidasari (Universitty of Salford) and Mr Aris Syntetos (University of Salford) The forecasting and stock control modules are the most important elements of an inventory system. The trade-off between customer service level and inventory investment is addressed in practice by formal quantitative inventory management (stock control) solutions. Given the very high number of Stock Keeping Units (SKUs), inventory management is typically fully automated (using software packages). However, the output of statistical software (such as demand forecasts or replenishment quantities) is often judgementally adjusted by managers. With regards to forecasting, empirical research suggests that practitioners rely heavily on judgemental forecasting methods. Further, when quantitative methods are used, they are very frequently judgementally adjusted. However, practitioners often adjust stock replenishment orders, not the forecasts. At the moment there isn’t a single empirical study that explores this area. The aim of our work is to explore the effects of incorporating human judgement into such inventory related decisions. Research propositions (research questions) are developed based on a literature review followed by exploratory empirical data analysis and simulation, and further validated through interviews. An extended database of 20,000 SKUs from the electronics industry is analysed for that purpose and insights are being offered on the performance of judgemental adjustments. 74 Making An Impact John Ranyard Ruth Kaufman Mark Roper Organisers: John Ranyard, Ruth Kaufman and Mark Roper 05/09/2012 : 08:20-08:30 : JM Duddingston Code: OR54A718 MAI - Introduction and Welcome John Ranyard, Ruth Kaufman and Mark Roper The 'Making an Impact' stream is specifically designed for practitioners - people whose day-job is to help solve their organisation's problems, whether as in-house or external consultant. Each event within it is intended to help participants build on existing experience, develop practice, make links and enable crossfertilisation through sharing knowledge and expertise. 05/09/2012 : 08:30 - 09:30 : JM Duddingston Code: OR54A719 Speed Networking Dr Sophie Carr (Bays Consulting) Speed Networking (facilitated by Sophie Carr) No matter what the event it can be hard to meet others working in similar fields to you. There are always old contacts to catch up with and revisit. Yet, now, more than ever your network of contacts is vital to helping you succeed ” whether that is developing new opportunities or learning about communities of practice. So, come along and start the MAI day really positively at the speed networking session. This structured, fun and focused session is specifically designed to help you meet as many people as possible in a series of short face to face meetings. Whether from an independent consultancy, large company or academia, come and develop links with others working in OR. At each face to face meeting both people will each have 2 minutes to present themselves ” talking about, for example, why they are at the conference, what areas of OR they specialise in or are interested in, what makes their work interesting. At the end of the meeting, you move onto the next one. In 60 minutes you will meet about 10 new contacts. The initial meeting will let you know how much you have in common ” the real value comes from following up the initial meeting and there is plenty of time at the conference to do so, maybe over coffee during the day or at the banquet? 05/09/2012 : 16:30-17:45 : JM Pentland Code: OR54A724 Practitioner – Academic Collaboration: Maximising Research Impact Chaired by Sally Brailsford (University of Southampton) The aim of this session is to explore how practitioners can get the most value out of OR research and help academics to maximise the real world impact of their work. Many practitioners are keen to utilise the latest developments to improve their effectiveness but face all sorts of barriers, real and perceived. From the academic point of view, it is well known that the next REF will include impact as a key criterion, and much discussion has been held about this; but do academics and practitioners have the same interpretation of ‘impact’? Prof Sally Brailsford will chair this interactive session, which will commence with brief introductory comments from a distinguished panel, comprising two experienced practitioners (Sean Jones, NATS and 75 Tony O’Connor, GORS); and two senior academics (Prof Mike Pidd, Lancaster, and Prof Stewart Robinson, Loughborough), who will briefly present their personal perspectives and give some innovative suggestions for how these barriers may be overcome. A general discussion will follow. 76 Making an Impact – Technique Tasters and Facilitated Workshops on ‘Hot’ Topics in O.R. Ruth Kaufman Kate Swatridge Organisers: Ruth Kaufman and Kate Swatridge Technique Tasters – Morning Sessions 05/09/2012 : 11:00-11:45 : JM Salisbury Code: OR54A763 Technique Taster - Rich Pictures Dr Giles Hindle (University of Hull) All O.R. projects require the project team to appreciate a complex problem situation. However bad an artist you are, this technique provides a mechanism for learning about the situation and agreeing a project specification. The technique can be employed individually or in groups and used by analysts or inexperienced participants. 05/09/2012 : 11:00-11:45 : JM Pentland West Code: OR54A765 Technique Taster - Bayesian Methods Dr Sophie Carr (Bays Consulting) Uncertainty in your numbers? Don’t know what to do? Don’t ask the audience or go 50:50. Phone a friend! Use your friend Bayesian Analysis! This whistlestop tour will show you how uncertainty can count and should not be feared. 05/09/2012 : 11:45-12:30 : JM Pentland West Code: OR54A766 Technique Taster - Constraint Programming Mr Guido Diepen (AIMMS Professional Services) Constraint Programming is a mathematical programming approach that enables the capturing of many structures observed in reality in a very natural manner; it has applications in scheduling, planning and routing amongst others. 05/09/2012 : 11:45 -12:30: JM Salisbury Code: OR54A764 Technique Taster - Systems Modelling Dr Giles Hindle (University of Hull) The systems modelling language from Soft Systems Methodology is flexible, powerful and a bit of a hidden gem. Discover this technique with expert facilitator Giles Hindle and come away ready to have a go in your own organisation. 77 Technique Tasters – Afternoon Sessions 05/09/2012 : 14:45-15:30 : JM Pentland Code: OR54A774 Technique Taster - Rich Pictures Dr Giles Hindle (University of Hull) All O.R. projects require the project team to appreciate a complex problem situation. However bad an artist you are, this technique provides a mechanism for learning about the situation and agreeing a project specification. The technique can be employed individually or in groups and used by analysts or inexperienced participants. 05/09/2012 : 14:45-15:30: JM Prestonfield Code: OR54A772 Technique Taster - Bayesian Methods Dr Sophie Carr (Bays Consulting) Uncertainty in your numbers? Don’t know what to do? Don’t ask the audience or go 50:50. Phone a friend! Use your friend Bayesian Analysis! This whistlestop tour will show you how uncertainty can count and should not be feared. 05/09/2012 : 15:30-16:15 : JM Pentland Code: OR54A773 Technique Taster - Systems Modelling Dr Giles Hindle (University of Hull) The systems modelling language from Soft Systems Methodology is flexible, powerful and a bit of a hidden gem. Discover this technique with expert facilitator Giles Hindle and come away ready to have a go in your own organisation. 05/09/2012 : 15:30-16:15: JM Prestonfield Code: OR54A771 Technique Taster - Constraint Programming Mr Guido Diepen (AIMMS Professional Services) Constraint Programming is a mathematical programming approach that enables the capturing of many structures observed in reality in a very natural manner; it has applications in scheduling, planning and routing amongst others. Facilitated Workshops on ‘Hot’ Topics in O.R. – Morning Sessions 05/09/2012 : 11:00 : JM Pentland East Code: OR54A770 Personal and Professional Development Ms Kate Swatridge (Kate Swatridge Consulting Ltd) Continuing development is both valuable and satisfying ” but not always easy to fit aound the day job, or to undertake on a limit ” or non-existent - budget. Bring ideas, share ideas and create your own development plan. 05/09/2012 : 11:00 : ST.L.Bonnar Code: OR54A767 Networking for Introverts Mrs Frances Sneddon (SIMUL8 Corporation) If you hate talking to strangers, instinctively dislike the idea of talking to somebody just to build your own network, can’t imagine why anybody would want to talk to you, but would like to connect with other professionals like you ” this is the place to come! Share tips, discuss difficulties, devise strategies for building a better network - and have a go at putting some of them into practice. 78 05/09/2012 : 11:00 : ST.L.Cowan Code: OR54A768 Build Your Own Pub! Mr Dave Buxton (Decision Lab) A chance to try out (or brush up on) your rapid-simulation skills and branch out into agent-based simulation. Have a go at building & running your own (simulated) pub where the drinkers make their own decisions! Competitive or collaborative ” it’s up to you! 05/09/2012 : 11:00 : ST.L.Brewster Code: OR54A769 Data Visualisation Mrs Jacqui Taylor (FlyingBinary Limited) Data visualisation is becoming indispensable to help organisations understand their own data and communicate it to others. Explore some of the key techniques and see how they could be useful in your own organisation ” or even essential, as younger generations with predominantly visual ways of understanding the world enter the ranks of the decision-makers. Facilitated Workshops on ‘Hot’ Topics in O.R. – Afternoon Sessions 05/09/2012 : 15:00 : JM Salisbury Code: OR54A775 Build Your Own Pub! Mr David Buxton (Decision Lab) A chance to try out (or brush up on) your rapid-simulation skills and branch out into agent-based simulation. Have a go at building & running your own (simulated) pub where the drinkers make their own decisions! Competitive or collaborative ” it’s up to you! 05/09/2012 : 15:00 : JM Holyrood Code: OR54A776 Data Visualisation Mrs Jacqui Taylor (FlyingBinary Limited) Data visualisation is becoming indispensable to help organisations understand their own data and communicate it to others. Explore some of the key techniques and see how they could be useful in your own organisation ” or even essential, as younger generations with predominantly visual ways of understanding the world enter the ranks of the decision-makers. 05/09/2012 : 15:00 : JM Duddingston Code: OR54A777 Networking for Introverts Mrs Frances Sneddon (SIMUL8 Corporation) If you hate talking to strangers, instinctively dislike the idea of talking to somebody just to build your own network, can’t imagine why anybody would want to talk to you, but would like to connect with other professionals like you ” this is the place to come! Share tips, discuss difficulties, devise strategies for building a better network - and have a go at putting some of them into practice. A Workshop not scheduled in the afternoon – but may run depending on demand: Personal and Professional Development Ms Kate Swatridge (Kate Swatridge Consulting Ltd) Continuing development is both valuable and satisfying ” but not always easy to fit around the day job, or to undertake on a limit ” or non-existent - budget. Bring ideas, share ideas and create your own development plan. 79 Additional Workshops that may run in the morning and afternoon, depending on demand are: Horizon Scanning: Sharing Experiences: Many of us undertake some form of horizon scanning, whether a quick mental check of what threats, opportunities and changes might be approaching, or a fullscale tour of the horizon. Here’s an opportunity to share experiences and thinking. Ethical Dilemmas in O.R.: ever been tempted to keep quiet about a mistake you’ve found in your numbers? or to go along with a forceful client instead of challenging them? This is a repeat of the workshop that received the highest satisfaction rating at last year’s conference. Kate Swatridge, decisionLab Selling O.R.: How do you persuade someone that you are the answer to their problems? Almost everybody faces that issue at some point in an O.R. career; what can we learn from each other? 80 Metaheuristics Andrew Parkes Ender Ӧzcan Organisers: Ender Ozcan and Andrew Parkes 04/09/2012 : 13:30 : ST.L.Cowan Code: OR54A548 Variable Neighbourhood Structures Dr John Lamb (University of Aberdeen) Meta-heuristics often allow search over different neighbourhoods of an incumbent solution S. A variable neighbourhood structure is a set of neighbourhoods of S that, in some measure, get further and further from S. Typically, a method such as variable neighbourhood search, might search thoroughly in neighbourhoods near S and use neighbourhoods further away for avoiding local minima. A variable neighbourhood structure that has pairwise disjoint neighbourhoods whose union is the set of all solutions is clearly preferable to one where a solution may occur in several of the neighbourhoods or none. Call such a neighbourhood partitioning. In practice, variable neighbourhood structures are often defined ad hoc according to the operations conveniently available to construct them and so are not always partitioning. This article shows that it is often possible to define a partitioning variable neighbourhood structure. It shows how such structures can be defined when the symmetric difference of two solutions is meaningfully defined, In particular it finds partitioning variable neighbourhoods for cycles and circuits and shows how an understanding of neighbourhood structures helps in designing better meta-heuristics. 04/09/2012 : 14:00 : ST.L.Cowan Code: OR54A625 Heuristic Function Resynthesis via Algebraic Machine Learning Dr Jerry Swan (University of Stirling), Prof Edmund Burke (University of Stirling), Prof Graham Kendall (University of Nottingham in Malaysia) and Dr Ender Ozcan (University of Nottingham) Permutation problems such as the n-puzzle have structure that can be captured algebraically. In particular, heuristic functions for these problems can be described in terms of the symmetric group. By employing techniques from the nascent discipline of algebraic machine learning, we investigate the harmonic decomposition of these heuristics via the extension of the Fourier transform to non-commutative groups. The Fourier-space representation for these heuristics can be manipulated via elements of a real-valued weight space of dimension equal to the number of conjugacy classes of the group. By applying a weighting followed by the inverse transform, we may obtain a `filtered' version of the original heuristic. Using the npuzzle as a case study, we show that the heuristic functions given by the Hamming and Manhattan distance metrics on puzzle state can be transformed into significantly better heuristics via a search in the weight space. 04/09/2012 : 14:30 : ST.L.Cowan Code: OR54A680 An Instance-specific Parameter Tuning Strategy (IPTS) in Practice: An Example of a German Temporary Employment Agency in the Construction Industry Dr Jana Ries (University of Portsmouth) and Dr Patrick Beullens (University of Southampton) Temporary employment agencies (TEA) provide workforce solutions to companies in various industries. The business strategy relies on contractors who request a number of employees for a set period of time. In the 81 particular case of the construction industry, employees are usually scheduled to work on various construction sites in different locations. Hence, a workforce is periodically sent to different locations for a particular period of time. The challenge of a TEA is to decide which set of employees (team) to send to which location based on individual qualification, transport costs and subsistence costs for each employee. In the presented case study, the workforce of the TEA has the option to use a transport vehicle of the company-owned fleet. In addition, the current TEA strategy allows employees to keep company-owned vehicles over the weekend at home if this is more efficient than returning the vehicle to the head office. Alternatively, employees are allowed to use their own private vehicle. Based on those options, routes are planned according to dynamically changing depots. Structural instance information include, for example, number of building sites, number of employees and spatial information of the area in which company depots, building sites and employee locations are situated. The problem is solved using a TabuSearch approach with an incorporated instance-specific parameter tuning strategy (IPTS) which allows to take into consideration structural information and a decision maker preference on heuristic performance in order to calibrate TabuSearch. The IPTS design uses fuzzy logic and features an automated set-up to facilitate its use for non-experts. 05/09/2012 : 08:30 : ST.L.Cowan Code: OR54A697 A Hyper-heuristic Approach to Construct Constrained-version of Magic Squares Mr Ahmed Kheiri (University of Nottingham) and Dr Ender Ozcan (University of Nottingham) A magic square is a square matrix of size (order) n contains distinct entries from 1 to n*n such that the summation of the entries along each row, column and diagonal have the same constant value. Constructing a magic square of a given order is considered as a computationally difficult permutation problem, particularly when additional constraints are imposed. Hyper-heuristics are high-level general search methodologies which operate on the heuristics rather than solutions, directly. As an extension to our previous work, we present an effective and efficient selection hyper-heuristic for constructing the constrained-version of magic squares. 05/09/2012 : 09:00 : ST.L.Cowan Code: OR54A710 A Hybrid Genetic Algorithm for the Travelling Repairman Problem Miss Cagla Cergibozan (Dokuz Eylul University) and Dr A. Serdar Tasan (Dokuz Eylul University) Despite the fact that cost minimisation is the common objective of product and service suppliers, the customer oriented view is of vital importance in some cases. At this point, the Travelling Repairman Problem (TRP) is examined to ensure customer satisfaction. TRP is the problem of finding a Hamiltonian cycle in which the objective is to minimise total waiting time of all customers that are situated at different locations. In the literature, this problem is also known with other names such as the Minimum Latency Problem (MLP) and the Deliveryman Problem (DMP). The TRP is a NP-hard problem; several exact solution algorithms, approximation algorithms and also meta-heuristic approaches for TRP are studied in the literature. The aim of this study is to develop an efficient and effective algorithm that can be applicable to real life problems. In this study, a hybrid algorithm which combines genetic algorithm with a local search heuristic is proposed to solve TRP. The algorithm is initialised with a random generated population and fitness values of candidate solutions are calculated by considering the features of a minimisation problem. Roulette wheel selection is used as selection procedure, order crossover is used as the crossover operator and two-exchange operation is used as a mutation operator. After mutation operation, two-opt local search heuristic is implemented according to a probability. Best parameter combination was determined by a set of computational experiments. The proposed algorithm was applied to a set of instances that have been studied in the literature and the performance of the approach was evaluated according to the results of the computational study. 82 Operations-Design-Innovation Interface Organiser: Shu-Jung Sunny Yang 04/09/2012 : 11:00 : ST.L.Brewster Code: OR54A654 Design and Operations for Online-Offline-Integrated Healthcare Network Facilitators Dr Jiun-Yu Yu (National Taiwan University) Healthcare services are now facing several challenges such as soaring costs and inappropriate allocation of resources. To respond to these challenges, the well-known management guru Clayton Christensen proposed three types of business models for healthcare service innovation: Solution Shops, Value-Adding Process Business, and Facilitated Networks. Although the impact of the first two types of business models are well discussed in the literature, the managerial implication of the third one is generally overlooked. In this study, organisations that facilitate interactions and resource exchanges among their participating entities in the healthcare system are referred to as ‚healthcare network facilitators‛. The website PatientsLikeMe and Taiwan Breast Cancer Alliance are two examples. Offline (physical) network business models, such as rail transportation and postal services, have been operating successfully long before the emergence of the internet, on which numerous online (virtual) business models are developed inspiringly in the past decade. Several attempts at developing innovative pure online healthcare network facilitators have been made since the beginning of the new century, but only a few of them succeeded. On the other hand, the advantage and power of virtual community seems not well understood by traditional offline healthcare network facilitators. This study aims to investigate the design and operations for online-offline-integrated healthcare network facilitators so that innovative healthcare services that are unlikely to be delivered in the past can be enabled now. A framework for such integration is to be proposed. 04/09/2012 : 11:30 : ST.L.Brewster Code: OR54A657 Modelling Alternate Bargaining using Target, Reserve and Minimum Price Thresholds Dr Andrew Brint (Sheffield University), Dr Andrea Genovese (Sheffield University Management School) and Dr Niraj Kumar (Sheffield University Management School) Alternate bargaining is where a buyer and a seller take turns in proposing a transaction price, with negotiations continuing until either an agreement is reached or the negotiations break down - common examples are house purchases (in England and Wales) and car purchases. Existing models assume that the parties behave rationally with the impetus to reach an agreement stemming from the threat of the appearance of a third party who might then replace one of the participants. We believe that behavioural factors are also an important part of the negotiations and this talk looks at modelling the bargaining process using three prices for each party: an optimistic price they hope to achieve, a basic "break even" price and a price at which negotiations break down if the offer does not meet it. 06/09/2012 : 08:30 : ST.L.Cowan Code: OR54A674 Hospital-acquired Infection Control with Strategic Healthcare Workers Miss Wenlin Chen (University of New South Wales), Prof Chung-Li Tseng (University of New South Wales) and Dr Shu-Jung Sunny Yang (University of Essex) 83 Although hospitals require healthcare workers (HCWs) to wash hands in every hand hygiene opportunity, healthcare literature points out that HCWs decide hand hygiene adherence strategically. We develop a biform game-theoretic model that integrates infection transmission dynamics and HCWs’ strategic hand hygiene behaviour. Our analysis helps design an effective hospitals’ infection control strategy to minimise the prevalence of antibiotic-resistant bacteria and reduce the likelihood of HCWs opportunistic behaviour. 06/09/2012 : 09:00 : ST.L.Cowan Code: OR54A681 Schumpeterian Entrepreneurship: An Adaptive Dynamics Approach Dr Yanto Chandra (University of Leeds) and Dr Shu-Jung Sunny Yang (University of Essex) We develop an adaptive dynamics model to investigate when strategies related to non-predictive control, such as effectuation, lead to stronger product market performance than prediction-based strategies, such as causation. The model yields distinctive implications about how to launch creative destruction from start-ups perspective and how to avoid creative destruction from incumbent perspective. Based on these implications, testable propositions are offered as to how trade-offs between non-predictive and predictive control in firmlevel strategic decision making depends on not only competitive dynamics between competing firms but also market dynamics involving the evolution of consumer types. 06/09/2012 : 11:30 : ST.L.Cowan Code: OR54A672 The Impact of Competition on Firm-level Capability Development Path Miss Yan Emma Liu (University of Essex) and Dr Shu-Jung Sunny Yang (University of Essex) Extending competitive dynamics perspective to the context of process improvement and innovation, this study investigates whether and under which conditions firms may benefit by launching a revolutionary process deriving from a well-established process. We develop insights regarding process competition through the application of a dynamic, computational model. The model draws on the history of process innovation in car manufacturing and the underlying theoretical relationships in the empirical literature. The results of our simulation experiments show that competition not only increases the frequency of process innovation for long-term growth but also decrease it for short-term survival. Testable propositions are offered as to how trade-offs between improvement and innovation in process management depends on external and internal competitive tensions. 06/09/2012 : 12:00 : ST.L.Cowan Code: OR54A741 Effectiveness of Complex Supply Network Information-sharing and Coordination Strategies: Implications of Hierarchy Mrs Mohsen Jafari Songhori (The University of Melbourne), Dr Alan Smith (The University of Melbourne) and Dr Sunny Yang (University of Essex) During the last few years, complex Supply Networks (SNs) have been developing as a new field in the Supply Chain and Operations Management area with at least two distinct perspectives. Both theoretical complex theory and empirical evidences assert the emergence of the partial hierarchical structures among firms in complex industries and SNs. Thus, firms might adopt high or low levels of hierarchy in SNs. Also, according to the organisational contingency and SCs theories, firms in such complex SNs need to select appropriate information sharing and coordination strategies to match (fit with) the imposed environmental uncertainty from their trading partners (e.g. suppliers, customers). Under this latter perspective, the strategies of firms themselves are sources of uncertainty for their trading partners. Thus, SN firms’ information sharing and coordination strategies might reduce their own uncertainty but create uncertainty for their trading partners. In this paper, a new perspective and model of complex SNs is developed to investigate differences between high and low level hierarchy firms’ uncertainty reduction. The results reveal that the low and middle hierarchy-level firms are more prone to ineffectiveness of their uncertainty reduction strategies. 84 06/09/2012 : 13:30 : ST.L.Cowan Code: OR54A670 Altruistic Entrepreneurship: Formal Modelling and Evolutionary Psychology Theorising Dr Yanto Chandra (University of Leeds), Miss Yan Emma Liu (University of Essex) and Dr Shu-Jung Sunny Yang (University of Essex) We examine the potential benefits of reciprocal altruism to taking entrepreneurial actions in supply chain recovery. Entrepreneurship literature suggests that entrepreneurial firms always demonstrate a strong preference of independence when pursuing potential opportunities in uncertainty. But in condition of sustaining collective firm performance, such as supply chains, anecdotal evidence indicates that altruism works well to take advantage of firm interdependency and quickly spread individual opportunities to the related others. Grounded on empirical observations, a dynamic, computational model is created to portray the dynamic evolution of operational performance among supply chain parties under the threat of unanticipated disruptions including natural and man-made disasters. By carefully chosen computational analysis, we offer testable propositions regarding altruistic entrepreneurial actions for supply chain recovery facing major disruptions. Our findings are translated into a managerial framework for formulating entrepreneurial strategy for recovering from disasters. 06/09/2012 : 14:00 : ST.L.Cowan Code: OR54A677 Shelf-space Competition: The Role of Store Brands Dr Chia-Wei Kuo (National Taiwan University) and Dr Shu-Jung Sunny Yang (University of Essex) We analyse the impact of the shelf-space level decision on managing channel conflict between store and national brands. Most previous work on the resulting channel interactions assumes that national brand manufacturers always get enough shelf-space for their products. However, this assumption is problematic since the category shelf-space level is determined by the retailers; so we relax it in this study. In our retail supply chain setting, the retailer sells its own store brand product along with a competing national brand in a consumer good category. Both brands compete for the scarce shelf-space. As a result, the most likely equilibrium outcome is that the available selling amount of each brand is constrained by the shelf-space available for its products and both brands coexist in the category. We demonstrate in our analytical model that the retail price of the store brand is increasing in its market share but decreasing in slotting allowance for renting the national brand shelf-space. Furthermore, the retail price, wholesale price, and shelf-space of the national brand are increasing in the category market size. 06/09/2012 : 14:30 : ST.L.Cowan Code: OR54A653 The Impact of Aging on Operations Re-design and Service Innovation of Medical Institutes Dr Jiun-Yu Yu (National Taiwan University) The speeding aging trend in Taiwan has triggered huge concern about the demand and supply of medical and healthcare services from every aspect of the society. It is expected that the needs for medical services related to chronic, long term, and aging issues will substantially increase in the coming decade. However, the healthcare services provided by current medical institutes seem not capable of meeting such growth in demand. Thus, innovative healthcare services and the corresponding service operations will have to be developed from an integrated and forward-looking perspective. On the other hand, the healthcare service providers, i.e. doctors and nurses, are aging as well. This generally-overlooked fact suggests that, to maintain the current medical service quality and efficiency, the existing healthcare practices and the corresponding service operations and organisational structures will have to be re-designed. With aging patients and doctors, healthcare service providers from different departments will be expected to collaborate more frequently, and the latest information and communication technology (ICT) products and services are expected to play a more important role in this re-design and innovation process. This study aims to apply the constructs from operations strategy to explore the potential thinking directions for the medical institutes to develop innovative healthcare services and to re-design their service operations. Decisions such as capacity, business processes, and organisational structure are to be investigated in particular. 85 Optimisation Organiser: Konstantinos Kaparis 04/09/2012 : 13:30 : 4 JM Holyrood Code: OR54A588 Solving Constraint Programming Problems with AIMMS Dr Guido Diepen (AIMMS) AIMMS is a mathematical modelling language and besides the support for the traditional mathematical programming like (non-)linear programs and mixed integer programs, AIMMS now also has support for constraint programming (CP). A big advantage of CP is that it allows for a lot more freedom in your modelling. Looking at the type of problems that you can model with CP, it typically performs very well on highly combinatorial problems, which normally are difficult to model/solve with mathematical programming. AIMMS has support for the most common global constraints from CP like for example the count constraint or the alldifferent constraint. Besides these CP essentials, AIMMS also has support for some specific scheduling constraints based on the notion of activities and resources where activities need a certain amount of the resources to finish. Additionally, sequencing of the activities can be modelled easily. These scheduling constraints are very useful when modelling any scheduling problem with CP in AIMMS. In this talk, I will explain some of the basics behind constraint programming and I will show how this new modelling concept can be used within the AIMMS modelling environment. 04/09/2012 : 14:00 : JM Holyrood Code: OR54A685 Stochastic Optimisation of Reliability Development Programmes Dr Kerem Akartunali (University of Strathclyde) and Prof John Quigley (University of Strathclyde) During product development a design will be subject to detailed analysis to improve reliability by identifying and removing faults. Reliability analysis will include tasks such as fault tree analysis, physics-of-failure analysis, and accelerated testing. Many tasks can be resource intensive and each may have varying degrees of efficacy. There is a need to plan an effective and efficient reliability program, but achieving this is nontrivial because each task may expose several different faults and each fault may be exposed by several different tasks. Thus the marginal benefit of introducing new analysis tasks within a reliability program will eventually be outweighed by the marginal cost of implementing the analysis. In this paper we use an existing model to anticipate the reliability of a product following a series of development activities. The model is stochastic in so far as the benefits in terms of reliability following activities are not deterministic but random. We develop a stochastic programming structure for this model to identify the sequence of activities that will minimise the expected programme costs. 04/09/2012 : 14:30 : JM Holyrood Code: OR54A714 A Great Deluge Based Learning Hyper-heuristic for Multi-objective Optimisation Miss Mashael Maashi (University of Nottingham), Prof Graham Kendall (University of Nottingham) and Dr Ender Ozcan (University of Nottingham) Hyper-heuristics have drawn increasing attention from the research community in recent years, although their roots can be traced back to the 1960s. They perform a search over the space of heuristics rather than searching over the solution space directly. In the hyper-heuristic framework, there are two main stages: 86 heuristic selection and move acceptance. In general, the movement acceptance can be deterministic or nondeterministic. In this study, we propose an extended choice function based hyper-heuristic for multiobjective optimisation based on the great deluge algorithm (GDA). In the original approach, the choice function control and manages a set of low level heuristics and aims to choose the best heuristic at any given time using historic performance to make this decision, along with the need to diversify the search at certain times. All Move is utilized as a deterministic move acceptance strategy meaning that we accept the output of each low level heuristic whether it improves the quality of the solution or not. In this paper, we employ a non-deterministic move acceptance strategy using the great deluge algorithm that accepts only improving moves and worsen moves in limited spaces under the boundary condition. As our hyper-heuristic approach is designed for multi-objective optimisation, D metric integrated with the GDA as a comparison tool. The rain speed parameter in GDA is set to different values in order to investigate the effectiveness of this parameter on the quality of solutions. The experimental results demonstrate the proposed approaches with different speed rain setting perform better than the original approach, and it shows that the speed rain settings are highly problems depended. All hyper-heuristic approaches are tested on the Walking Fish Group test suite, a common benchmark for multi-objective optimisation. 04/09/2012 : 15:30 : JM Holyrood Code: OR54A571 Combination Booking Models for Air Cargo Revenue Management Miss Emily Cookson (Lancaster University), Prof Kevin Glazebrook (Lancaster University) and Prof Joern Meissner (Kühne Logistics University) Air cargo revenue management can be described as the integrated management of aircraft cargo weight and volume constraints, and chargeable cargo rates. The Air Cargo industry faces some unique challenges: Highly volatile demand, weight/volume uncertainty through variable tendering and very short booking cycles are some of the features. Booking requests can be divided into long-term contract (allotment) and short-term ad-hoc (spot-sale) types which have significantly different characteristics. In the case of combined passenger and cargo air carriers, the prioritisation of passengers and their baggage further challenges the allocation of aircraft belly space. With the current economic climate and increasing fuel prices, airlines are faced with a mounting pressure to maximise revenue. We have developed a new dynamic programming formulation with potential to aid an air cargo company to maximise its expected profit under these challenges. The initial model focuses on the spot-sale booking process, making acceptance decisions using cargo prices formulated under uncertainty. Since the dynamic programming formulation is computationally intractable due to the large number of states, the non-linearity of the value functions, high dimensionality and the complexity of the integrals, a discretised version and approximate solutions have been created using numerical analysis methods, including spline interpolation and numerical integration over probability confidence intervals, to form an executable system in MATLAB with NAG and CPLEX. Escalating penalty functions have been included as an optional disincentive for extreme overbooking. The proposed model performs significantly better than several comparable policies. As an extension to our booking control mode, we investigate the proportion of long-term versus short-term contracts. The current research aims to create a strategic decision model that will integrate the allotment and spot-sale booking processes under uncertainty and whilst maximising expected profits. We report on our first results. 04/09/2012 : 16:00 : JM Holyrood Code: OR54A596 On a Warm-started Primal-dual Column Generation Method Mr Pablo Gonzalez-Brevis (University of Edinburgh) and Prof Jacek Gondzio (University of Edinburgh) In this talk the primal-dual column generation method will be presented. The method is based on interior points and relies on sub-optimal and well-centred solutions of the restricted master problems. Although it uses sub-optimal solutions the method still finds the optimal solution of the master problem. Additionally, warm-starting strategies in this context will be discussed. Numerical experiments comparing the method with the standard column generation and the analytic centre cutting plane method will be presented. These 87 results show the benefits of using this warm-started primal-dual column generation method when solving the relaxation of three well-known combinatorial optimisation problems after applying Dantzig-Wolfe decomposition. 04/09/2012 : 16:30 : JM Holyrood Code: OR54A664 Radiation Treatment Planning Optimisation for Volumetric Modulated Arc Therapy (VMAT) Dr Kerem Akartunali (University of Strathclyde) and Dr Vicky Mak-Hau (Deakin University) Volumetric-Modulated Arc Therapy (VMAT) is a recent technological development in the area of cancer radiation treatment, with the aim of generating a treatment plan of appropriate radiation dosages, angles and collimator shapes. The problem is computationally very challenging, in particular considering additional problem features such as time limitations. In this talk, we will discuss an integer programming formulation for this problem, with improvements using some linearization techniques. Then we will discuss a branching and column generation framework specifically designed for this problem. We will present some computational results, and conclude with possible future extensions. 05/09/2012 : 09:00 : ST.L.Bonnar Code: OR54A641 A Generalized Returns-to-scale Model in DEA Dr Miki Tsutsui (Central Research Institute of Electric Power Industry) and Prof Kaoru Tone (National Graduate Institute for Policy Studies) In data envelopment analysis (DEA), there are typically four types of returns-to-scale (RTS): constant (CRS), variable (VRS), increasing (IRS) and decreasing (DRS). They are classified by a constraint which defines the lower and upper bounds (L&U) of sum of components in the intensity vector (IV) in the production possibility set, i.e. L=0&U=infinity for CRS, L=1&U=infinity for IRS, L=0&U=1 for DRS, and L=1&U=1 for VRS. Originally in economics, ‚variable‛ returns-to-scale is the general case of RTS assumption, and thus CRS, IRS, and DRS are special cases of VRS. However in the DEA field, VRS indicates only a specific case as IV=1, represented by the BCC model. This model causes a problem that if a DMU has a minimum input value of all DMUs for any input item, or a maximum output value of all DMUs for any output item, the DMU is definitely evaluated as VRS-efficient. In many instances, a DMU with a very small score under the CRS model turns out to be full VRS-efficient by this rule. This phenomenon looks strange and many practitioners doubt this property. In order to prevent this VRS-DEA problem, in the current study, we focus on Generalised RTS (GRS) with constraint of the lower and upper bounds (L&U) of IV. This is more general model similar to VRS in Economics, i.e. CRS, IRS, DRS, and also VRS in DEA are the special cases of GRS model. We propose a practical method for determining L and U of IV. Furthermore, we define the scale efficiency score based on GRS instead of VRS, and identify the RTS type in order to clarify the source of the scale inefficiency. We also propose a method to obtain consistent RTS judgment even if we utilise multiple outputs. 05/09/2012 : 11:00 : JM Duddingston Code: OR54A577 Deployment of Multi-tier ICT Services in Cloud Data Centres Prof Bjorn Nygreen (Norwegian University of Science and Technology) and Mr Anders Nordby Gullhav (Norwegian University of Science and Technology) The Software-as-a-Service (SaaS) model of the cloud computing paradigm has changed the way ICT services are being provisioned, as the service providers are given more flexibility in scaling their services according to the demand and requirements of the end-users. In the provisioning, a SaaS provider needs to focus on cost- and energy-efficient operation of its private cloud, and ensure that the services deployed on the nodes in the private cloud, or on the virtual machines in a public cloud, have a Quality of Service (QoS) satisfying the agreed requirements. In this talk, we mainly focus on decisions of a SaaS provider related to the management of his services and cloud data centres, but also acknowledge decisions related to deploying services into public clouds. The services are modelled as consisting of multiple tiers, where each of these tiers is distinct software components with different functionality. Increased QoS, measured as service 88 response time and availability, is obtained by adding active or passive copies of the components in the various tiers, thereby providing more resources to the services. Modelling of multi-tier services are generally more complex compared to services of a single tier, since one cannot treat the tiers separately, and thus, there might exist several ways to satisfy the QoS requirements of a service. We will discuss different (M)IP models of the problem where the goal is to minimise the cost of running services in private and public clouds. 05/09/2012 : 11:30 : JM Duddingston Code: OR54A603 Matrix-free Interior Point Method for Sparse Signal Reconstruction Problems Mr Kimon Fountoulakis (The University of Edinburgh), Prof Jacek Gondzio (The University of Edinburgh) and Dr Pavel Zhlobich (The University of Edinburgh) We introduce a specialised version of an Interior Point Method (IPM) for sparse signal reconstruction problems. The proposed method is a Matrix-Free variant of IPM which takes full advantage of lowcomplexity matrix-vector product operations involved in sparse signal reconstruction. The particular features of constraint matrices in these problems such as orthogonality of rows and the Restricted Isometry Property (RIP) (which guarantees near-orthogonality of certain subsets of columns of these matrices) are exploited to design pre-conditioners for the KKT systems arising in IPMs. Applicability of the proposed pre-conditioners is justified with strong theoretical results and confirmed with numerical experiments. Results from a variety of large scale sparse signal reconstruction problems indicate that the Matrix-free IPM compares favourably with other state-of-the-art solvers. 05/09/2012 : 12:00 : JM Duddingston Code: OR54A729 Reliability and Maintenance of Complex Desalination Plant Subject to Failures Mr Ahmed AlHinai (Glasgow Caledonaian University) and Dr Babakalli Alkali (Glasgow Caledonian University) This paper presents a comprehensive failure mode and effect analysis conducted on critical equipment in a complex desalination plant. Classes of failure modes that contribute to equipment failure are identified. Historical failure and maintenance data are collected and analysed. We explored the failure data for trend and plotted graphs to have a good insight about the failures pattern using statistical approach. Graphical results show that the inter-arrival times between failures tend to be shorter at some points. The data exhibit random phenomenon and we assumed that the failure process follow a non-homogenous Poisson process. We use a standard statistical approach for reliability analysis by fitting the Weibull distribution to the data sets. The results obtained are presented on a distribution overview. This investigation set a pace for reliability modelling to schedule optimal cost effective preventive maintenance strategy for the desalination plant equipment. 06/09/2012 : 11:30 : ST.L.Bonnar Code: OR54A597 Dynamic DEA with Network Structure: A Slacks-based Measure Approach Prof Kaoru Tone (National Graduate Institute for Policy Studies) and Dr Miki Tsutsui (Central Research Institute of Electric Power Industry) We propose a dynamic DEA model involving network structure in each period within the framework of slacks-based measure approach. We have already published the network SBM (NSBM) and the dynamic SBM (DSBM) models separately. Hence, this article is a composite of these two models. Vertically, we deal with multiple divisions connected by links of network structure within each period and, horizontally, we combine the network structure by means of carry-over activities between two succeeding periods. This model can evaluate (1) the overall efficiency over the observed entire period, (2) dynamic change of period efficiency and (3) dynamic change of divisional efficiency. The model can be implemented in input-, outputor non-(both) oriented forms under the CRS or VRS assumptions on the production possibility set. We can impose the initial condition on the carry-overs. Non-uniqueness issue of term efficiencies is also discussed. 89 06/09/2012 : 13:30 : JM Pentland West Code: OR54A578 Medium-Term Trading Strategy of a Dominant Electricity Producer Mr Arash Mostajeran Gourtani (University of Southampton), Mr David Pozo (University of Castilla-La Mancha), Dr Maria Teresa Vespucci (University of Bergamo) and Dr Huifu Xu (University of Southampton) This paper presents a multi-objective two-stage bi-level stochastic programming framework for a dominant electricity producer to determine an optimal trading strategy in a deregulated electricity spot market in a medium-term time horizon. At the first stage and upper level, the dominant producer aims to maximise its expected market share and profit taking into account the trade-off between the two objectives, and at the second stage and lower level, the independent system operator (ISO) determines the dispatches and power flow on hourly basis after realisation of uncertainty in market demand by solving an optimisation problem which aims to maximise the total social welfare. Through Karush-Kuhn-Tucker conditions, the lower level problem is formulated as a complementarity problem and subsequently the dominant producer’s optimal decision making problem as a two stage Stochastic Mathematical Problem with Equilibrium Constraints (SMPEC).To solve the SMPEC, it is proposed to reformulate the SMPEC as a Mixed Integer Linear Program (MILP) by presenting the complementarity constraints as a system of mixed integer linear inequalities with binary variables. Numerical tests results are reported through an illustrative example and a medium size case study. 06/09/2012 : 14:00 : JM Pentland West Code: OR54A589 Cutting Planes for a Stochastic Variant of the Network Loading Problem Dr Konstantinos Kaparis (University of Southampton), Prof Adam Letchford (Lancaster University) and Prof Stein Wallace (Lancaster University) In the Network Loading Problem (NLP) one is given an undirected graph G=(V,E) along with a set of demands within V. The task is to choose the amount of integer capacities to be installed on each edge in such a way that the cost is minimised and all the commodities can be routed simultaneously. In this work, a stochastic variant of this problem is being considered. Moreover, the demand is uncertain and its realisation depends on a finite number of scenarios. We model this problem as a 2-stage stochastic mixed-integer program with complete recourse. We introduce the new class of probabilistic metric inequalities and we describe a heuristic separation algorithm. 90 O.R. & Strategy Frances O’Brien Martin Kunc Organisers: Frances O'Brien and Martin Kunc 04/09/2012 : 11:00 : ST.L.Nelson Code: OR54A524 Modeling Start-ups using System Dynamics: Towards a Generic Model Dr Martin Kunc (Warwick Business School) In system dynamics, different scholars have proposed templates to analyse similar problems, e.g. Peter Senge’s archetypes. However, there are not system dynamics archetypes of business models. This paper attempts to develop a generic model for managing start-ups based on two common businesses, internet services and management consultants, with the aim to provide a template for entrepreneurs to develop a system dynamics model of their own firm. The different experiments with the generic model suggest that long term management of staff and cost optimisation are the right strategies to overcome the high rate of failure during the first five years of operation of start-ups. 04/09/2012 : 11:30 : ST.L.Nelson Code: OR54A675 Competitive Archetypes: A Competitive Dynamics Perspective Miss Yan Emma Liu (University of Essex) and Dr Shu-Jung Sunny Yang (University of Essex) The study identifies and conceptualises the notion of competitive archetypes that are generic competition patterns in inter-firm rivalry. Generic patterns are highly effective in providing insights into the ‘nature’ of underlying problem structure and its associated behaviours over time. Yet among the leading approaches in strategy, little attention has been devoted to developing generic competition patterns. Inspired by the effectiveness of system archetypes in system dynamics, we adopt a systems perspective to develop the competitive archetypes to capture dynamic, competitive interaction in diverse settings of inter-firm rivalry and to increase our understanding of the nature of competition. Grounded on competitive dynamics research, the paper synthesises established constructs and embrace the cognitive frameworks into firm-level decision making process, proposing a set of competitive archetypes in inter-firm rivalry. 04/09/2012 : 12:00 : ST.L.Nelson Code: OR54A733 Supporting Internationalisation Strategies using System Dynamics Modelling: Lessons from a Strategic Development Approach Mr Juan Pablo Torres (The University of Warwick) The aim of this research is to show how modelling and simulation can support a deliberate strategy of internationalisation. I will show details and experiences of developing system dynamics models with five small companies to structure and evaluate their internationalisation strategy prior to implementation. From this process it emerged that the cognitive ability of the CEOs visualising the outcome of his/her strategic actions is limited due to the difficulty in measuring feedback processes conclusively during strategy design processes. Finally, I will present what happened to those companies after the modelling experience one year later. 91 04/09/2012 : 13:30 : ST.L.Nelson Code: OR54A644 Modelling Degradation of Control & Monitoring Assets using Bayes-Empirical Bayes Inference for a Time-Dependent Poisson Process Mr Graeme Blair (University of Strathclyde), Prof John Quigley (University of Strathclyde) and Prof Lesley Walls (University of Strathclyde) Scottish Water (SW) is a Government owned company which provides a vital service to the Scottish public. This includes the abstraction, treatment and consistent supply of clean safe drinking water to every home and business across Scotland. SW also collect, treat and dispose of wastewater in a manner that protects the environment. These processes are controlled and monitored throughout SW’s network of reservoirs, pipes, treatment works, and pumping stations. This has led to work, motivated by SW who seek, to model the deterioration of control and monitoring (CAM) assets to predict future failures. Model estimates will inform strategic investment budgeting and will be subject to scrutiny by the regulator. A non-homogeneous Poisson process (NHPP), with a power intensity function (often called a Power Law Process), provides a model for the degradation of CAM assets. For these assets there are limited historical statistical failure records. A novel approach is developed whereby the scale parameter of the Power Law Process is estimated using the available statistical data, but the shape parameter is estimated using structured expert judgement elicited from relevant engineers. Empirical Bayes inference provides an approach for estimating the frequency of rare failures of each class of CAM assets, as a weighted average of the frequencies of failure for that asset class and the failures in the pool of all CAM assets. In this problem, a hybrid of Bayes and Empirical Bayes is used to combine the subjective and empirical prior distributions for the shape and scale parameters, respectively. This talk describes the practical motivation and the application of the approach to a real problem. The method developed will be explained. The suitability of a Power Law Process as a model for degradation and the usefulness of empirical Bayes inference will be discussed. 04/09/2012 : 14:30 : ST.L.Nelson Code: OR54A659 Data Analysis in the Water Industry: A Good Practice Guide from an Application in Scottish Water Miss Efthalia Anagnostou (Scottish Water) and Miss Deborah Gee (Scottish Water) Data analysis is a modern day prerequisite to the optimised management of physical assets. In the UK water and wastewater industry, data-driven analytical methods now play a major role in the everyday decision making and strategic thinking of both the asset planning community and the market regulators. For example, the analysis of historical failure data is used to model the deterioration process of multiple assets in order to forecast likely changes in asset performance over time. Furthermore, as the value of analyticalbased methods has become more widely recognised, companies have increasingly diversified their businesses from predominantly engineering based organisations to information-centric organisations. With each company utilising the large quantities of data that is available to them alongside modern analytical methods to shape policy, improve customer service and reduce cost. Established in May 2009 as an affiliate of the Royal Statistical Society, the Water Statistics User Group (WSUG) aims to share statistical approaches and promote good-practice data analysis across the water and wastewater industry. The group currently includes representation from 26 water and wastewater companies. WSUG has developed a guide to good practice data analysis, which can be used by the industry participants and regulators to promote goodpractice and validate the quality of their analytical processes, models and outputs. It aims to encourage continual improvement in the standard of data analysis across the industry, ultimately driving companies to underpin key asset management decisions with consistently robust analytical results and data. The paper describes: (1) The data analysis spiral ” a flagship approach to data analysis comprising 7 key steps: capture stakeholder requirements, gather business data, conduct exploratory analysis, develop analysis, validate analysis, publish results and acceptance testing. (2) Examples of how the guide is applied within Scottish Water to benefit stakeholders within the business. 92 04/09/2012 : 15:30 : ST.L.Nelson Code: OR54A521 Is There Still Life in Stafford Beer's Viable System Model (VSM)? Dr Stephen Harwood (University of Edinburgh) It has been over 40 years since the Stafford Beer presented his Viable System Model (VSM). Whilst the VSM attracted the gaze of a wide audience during the latter part of the twentieth century, as a way to conceptualise the organisation of social systems, it has remained in the margins of academic discourse, with it perhaps disappearing from university syllabuses. It is proposed that the VSM does have a contribution with regard to current debates about organisations. One specific application relates to the strategy-aspractice movement with their concerns about how strategies emerge. There is a clear need to understand the complex interplay of routines, episodic events (e.g. meetings, workshops) and the role of discourse, from which strategies emerge, either as formal documents or as a series of aligned decisions. This raises questions about how to understand the complex interplay between the different levels within organisations, the relationships between policymakers and shop-floor practitioners, as well as with outside, and the distribution of discretion. It invites questions about organisational identity and how it is maintained or transformed over time. As regards debates specifically about routines, there is the issue of how routines are conceptually explained. Routines are not static but a dynamic feature of organisations which contribute to its stability. They can be viewed as one aspect of an organisation’s memory; the outcome of the accumulative experience of dealing with disturbances to the functioning of routines. Practices such as ‘lean’ invoke the stripping out of redundancy, but how is redundancy identified and is the outcome memory degeneration? It is proposed that the concepts of the VSM established by Stafford Beer and developed by others (e.g. Espejo, Espinosa, Harden) still offer a valid contributory insight which complement existing main stream thinking about organisations, particularly with its insights into recursion, autonomy and adaptation. 04/09/2012 : 16:00 : ST.L.Nelson Code: OR54A665 Putting the ‘Science of Change’ into the ‘Science of Better’ Mr Philip Jones (Defence Science and Technology Laboratory) This presentation will highlight the need for O.R. to embrace behavioural change science to improve the effectiveness of strategic O.R. interventions, whether they are intended to achieve improvements at the individual, organisational or societal level. It will: “ Summarise key behavioural change principles like Cialdini’s ‘Weapons of Influence’; the ‘Nudge’ technique; and, organisational and societal change success factors; “ Show how these approaches are being applied by the Ministry of Defence and other Government departments and illustrate their impact; and “ Argue the need for O.R. to make these ideas an integral part of the discipline if it is truly to be the ‘Science of Better’: working to deliver change in the real world, not just preferred options for change. The presentation is intended to be of interest to practitioners and academics alike, particularly those with in an interest in the strategic use of ‘Soft O.R.’ who are looking for avenues to take the area forward. 04/09/2012 : 16:30 : ST.L.Nelson Code: OR54A658 Reflecting on the Use of Social Media to Encourage Engagement with a Scenario Project to Explore the 'Future of Food' for Birmingham in 2050 Ms Frances O'Brien (University of Warwick) and Ms Maureen Meadows (The Open University) Scenario planning is a tool which captures multiple alternative possible futures that an organisation may face. The scenario planning setting is often the workshop. With the advent of social media it is possible for wider engagement in the process to be considered. This paper reports progress with an on-going project to develop scenarios for the future of the food system within the Birmingham ‘hinterland’ for 2050. It reflects on experiences using social media to encourage wider interest of and engagement with the project. 93 Code: OR54A639 05/09/2012 : 08:30 : ST.L.Nelson Group Dynamics and Intragroup Conflicts: A Mental Model Convergence Perspective Mr Kenneth Kyunghyun Huh (Warwick Business School) Past group process research suggests that having a high level of task related conflict while keeping emotional conflicts at a minimal level is the key to achieve an effective group outcome; yet the literature returns contradictory findings. This study supplements the theory by using the perspective of mental model convergence; a concept that is used to define a status where heterogeneous individual mental models converge into a shared mental model. A longitudinal study of six groups was conducted to investigate how intragroup conflict and various group dynamics are related with mental model convergence. 05/09/2012 : 09:00 : ST.L.Nelson Code: OR54A595 ‚Playing Seriously‛ with Strategic Management: A Case Study Matching Role Playing and System Dynamics to Develop SCM Policies Prof Federico Barnabè (University of Siena), Prof Cristiano Busco (University of Siena), Prof Pål I. Davidsen (University of Bergen), Mr Maurizio Lambri (GMV Consulting) and Mr Gianfranco Zatta (GMV Consulting) This work explores how role playing games may play a fundamental role in knowledge sharing and team learning, especially dealing with complex and dynamic business domains and facing strategic issues. When combined with a System Dynamics modelling intervention, this approach allows to support strategic thinking and learning in teams. Starting from the previous considerations, this work presents the main features and the outcomes of a specific case study. 05/09/2012 : 11:00 : ST.L.Nelson Code: OR54A535 Metaphorical Models and Credible Worlds: Insights from Small and Stylised Models Dr John Morecroft (London Business School) All models of business and society are stylised simplifications. It is therefore important to consider the credibility and value of model-based insights and to recognise why and when small models are appropriate and helpful. I probe these issues in the light of two classic models from the economics literature that have intrigued some theoretical economists. Drawing on this work I identify properties of small models that make them useful for inductive reasoning and learning. Such small and metaphorical models, though highly stylised, need to be both plausible and sufficiently understandable to stimulate comparisons with the real world. It is from such comparisons that inductive reasoning stems. I show similar properties in the wellknown World Dynamics model that addresses global industrial growth and sustainability. I demonstrate my argument with a selection of partial and full model simulations; each a credible world and each a metaphor for understanding global limits to growth and industrialisation. 05/09/2012 : 11:30 : ST.L.Nelson) Code: OR54A598 A Structural and Behavioural Triple Helix Spiral of Knowledge-producing Institutions Mr UngKyu Han (University of Warwick) and Dr Martin Kunc (University of Warwick) A triple helix thesis of university-industry-government has been more applied to analysing an innovation ecosystem. Yet, it is still difficult for academics and practitioners with divergent backgrounds to understand and share its dynamics since innovation is more than a linear path starting from R&D. To improve the communication between stakeholders of triple helix ecosystems, this paper is intended to expand the triple helix theories by means of systems thinking, reflecting spiral interactions of its structure and behaviours. In particular, we seek to account for knowledge-producing institutions - universities, industries, and government research institutes. To this end, we propose a causal loop diagram that illustrates complexity, nonlinearity, and recursion endogenous in a triple helix innovation system. The model comprises the missions of three institutions and the hybrid or intermediary organisations bridging the institutions. The proposed diagram helps in observing what is going on in multiple institutional initiatives across (intersphere circulation) and inside (intra-sphere circulation) the spheres entangled with the innovation 94 ecosystem. Moreover, it is useful to determine attitudes to newness of innovation during the decisionmaking processes on each institutional and policy scales, by recognising short- and long-term effects caused by intended actions throughout the complex, nonlinear, and recursive innovation process. Lastly, the combination of systems thinking and a triple helix thesis may make it more rational for innovation policy making processes to consider counterintuitive consequences generated in the innovation system. 05/09/2012 : 12:00 : ST.L.Nelson Code: OR54A586 Assessing Strategic Risk in the Public Sector: Is Incremental Change Enough? Mr Ian Birchmore (The University of Warwick) Local authorities have a legal duty to manage risk. The definition of risk and how it is to be assessed are rather less clear than might be expected given this legal duty. This paper arises from doctoral research into the assessment of strategic risk in the public sector. The research concentrates on risks in local authorities to draw on their diversity to provide a rich context. The paper will briefly set out and explore the nature of strategic risk in UK local authorities and propose a sense of strategic risk in the public sector that reflects its context and provides a base for assessment. It will then consider the ways in which local authorities have defined and implemented a matrix-based impact / likelihood risk assessment model, highlighting the limited strengths and greater weaknesses of these current approaches. The paper will then propose a series of changes to the current core approach which seek to maintain its accessible simplicity whilst addressing the identified fundamental weaknesses and will close by considering the question of whether such incremental change is enough. 95 O.R. Consultancy and Case Studies Organiser: Kuangyi Liu 06/09/2012 : 08:30 : ST.L.Nelson Code: OR54A678 Modelling the Net Benefits of Alternative Management Strategies for Bacterial KidneyDisease in Scottish Salmonid Aquaculture Mr John Soje (Marine Scotland - Scottish Government) Bacterial kidney disease (BKD) of farmed salmonid fish is caused by the pathogen Renibacterium salmoninarum which is often sub-clinical and might go undetected if diagnostic methods have low sensitivity. The observed prevalence of infected farms is low for salmon production relative to trout. Management measures such as movement controls constitute a regulatory burden, in financial terms, for both the Industry and Government. The awareness of the potential cost of disease in fish farming is increasing, but there has been little economic assessment of the costs of disease and benefits of intervention. This presentation describes a decision analysis model of the impact of BKD under different policy options. It uses event trees to assess the risks and then evaluate the net benefits of changing statutory controls, either strengthening or weakening them. In the tree, decision nodes depict policy options for which subsequent events have a probability of being uninfected or infected; this splits into diseased or sub clinically infected which in turn divides between those known or unknown to the authorities. These are states determined by a separate Susceptible-Infected (SI) model. These outcomes lead to consequent costs associated with on the one hand disease and on the other surveillance and movement restrictions. Total probabilities and costs can then be established at the end of each node; this however assumes that events are dependent. For each node the expected impact in any year as a result of changes in surveillance and intervention is the sum of the change in value from disease-related losses and the change in disease management costs to both business and the regulator. The model has helped the team to collectively structure the analysis and identify data sources from inspectorate, laboratory, and questionnaire etc. The approach can be used for other diseases. 06/09/2012 : 09:00 : ST.L.Nelson Code: OR54A682 SWOT Analysis using General Morphological Analysis: A Pharmaceutical Case Study for New Business Drivers Dr Nasir Hussain (Strategy Foresight LLP), Mr Bruce Garvey (Strategy Foresight LLP) and Dr Tom Ritchey (Swedish Morphological Society) The humble SWOT is a remarkably simple problem structuring framework. Since its introduction by Albert Humphrey in the ‘60s, it has surprisingly undergone little revision - the lack of facilitation of constructing a SWOT in the business setting and particularly the haphazard way of inter-relating the various elements within each quadrant fails to realise the technique’s true potential. In this paper, we report a major improvement of how a SWOT should be constructed and analysed by using the process of Cross Consistency Assessment (CCA). Here, each and every suggestion from each quadrant is compared pairwise to test for compatibility. The CCA is similar to a Cross Impact Analysis except that no directional or causal linkage is assumed but merely mutual consistency in the arguments. CCA is actually an essential element of General Morphological Analysis a problem- structuring method that permits the structuring and analysis of multidimensional problems. To test the validity of SWOT-MA™ process, an actual case example involving a start-up pharmaceutical firm was considered for the following focus question: ‚What are the most 96 important factors to secure drug approvals within funding time frame of 3 years‛. 22 factors were identified in the four parameters. This would necessitate that 840 unique configurations be considered, an onerous task for resource-constrained organisations (there were 279 pairs for the SWOT alone). An additional output parameter, the Ansoff’s Matrix, was then bolted onto the SWOT. This ‘5th column’ is an essential step to make sense of what one is trying to achieve with the SWOT-MA™ exercise. This second CCA iteration resulted in 23 unique configurations out of a possible 3360 combinations, a reduction of 99% of the entire problem space. We believe that such an approach can be applied to other commonly applied business management tools where multiple parameters must be considered e.g. PESTLE, VRIO. 06/09/2012 : 09:30 : ST.L.Nelson Code: OR54A687 Using Modelling to Facilitate Collaboration in the NHS Mr Nicholas Jones (PwC) In February 2012, PwC were commissioned to assess the options available to one of the most financially challenged NHS trusts in the country, while also considering the impact of any such changes on the wider health economy. Previous attempts to address this challenge had struggled to galvanise people’s thinking so we sought a different approach, combining collaborative engagement and focused system-wide modelling to identify clinically viable solutions that could close the financial gap. Working to very demanding timescales, we developed a model that allowed for estimation of the financial impacts associated with changes in service provision or commissioning patterns. This model formed an important part of a collaborative event involving key stakeholders (the trust in question, other nearby providers, clinical commissioning groups, the strategy health authority, etc.). In this way, the model was used to create broad engagement, stimulate thinking on options for change, challenge preconceptions and identify any scenarios that should be explored in more detail following the study. The study demonstrated that modelling is a powerful tool to facilitate system-wide thinking and option development, focusing stakeholders on changes that would make a material impact. It also indicated that a solution to the challenge was feasible and helped to establish momentum for change amongst the stakeholders. 06/09/2012 : 11:30 : ST.L.Nelson Code: OR54A601 Managing Projects in an Uncertain World: Engaging Stakeholders and Building a Systemic View of Risk Prof Fran Ackermann (Strathclyde University), Prof Susan Howick (Strathclyde University), Dr Tom Houghton (Strathclyde University), Prof John Quigley (Strathclyde University) and Prof Lesley Walls (Strathclyde University) As evidenced through both a historical record of project over-runs and a large number of projects currently under media spotlight for cost over-runs e.g. Edinburgh Tram system, managing projects can be a risky business. Managers are faced with effectively engaging and working with often a multitude of parties (suppliers, government, consultants, general public and own staff), anticipating and dealing with a wealth of often interlocking uncertainties (ranging from technological innovation to force majeure), and often undertaking these activities within a compressed timeframe (which exacerbates the situation) This paper describes a risk management process that has been developed to assist managers facing such situations. The process explicitly engages a range of stakeholders using a group decision support system (allowing for equality of view) and causal mapping process (which facilitates the surfacing of risks along with their ramifications), and thus provides not only a comprehensive appreciation of the totality of the risks identified but also a greater understanding of their subtleties, consequences etc. The paper will discuss a real case where the process was used to gain an understanding of the risks associated with developing a new energy system. The paper will describe the process and its implications, before reflecting on the insights, limitations and future research. 97 06/09/2012 : 12:00 : ST.L.Nelson Code: OR54A731 Prioritising Replacements of Protection Assets in an Electricity Distribution Network Ms Liz Archibald (Decision Lab) As part of a regulated sector, Distribution Network Owners (DNOs) in the UK electricity industry have to submit their investment plans for approval by the regulator OFGEM. In the latest guidance from OFGEM, DNOs are strongly advised to use condition or evidence-based analysis via appropriate modelling techniques for the justification of replacement expenditure. For some assets, data availability is limited, or its suitability for modelling may not be ideal. These practical issues can present significant challenges in the development of useful decision support tools. This case study describes the development of a modelling tool which transparently captures expert engineering judgements and uses these in combination with best available age and condition data to produce annual asset replacement levels for protection relays. Identification of an appropriate modelling platform, strong client engagement and a collaborative approach to the various data challenges were the key to the success of this consultancy project. 06/09/2012 : 13:30 : ST.L.Nelson Code: OR54A580 A Multi-methodology Approach to Comprehensive CB (RN) Risk Assessment and Mitigation Mr Diederik Wijnmalen (TNO) This presentation focuses on an application and further enhancement regarding CB(RN) (chemical, biological, radiological, nuclear) incidents of the OECD-acclaimed Dutch national risk assessment methodology. It combines morphological analysis, scenario development and analysis, risk analysis taking both likelihood and a broad spectrum of impact types (having a quantitative or an inherently qualitative nature) into account, multi-criteria analysis using weighting profiles based on human value systems, capability analysis and cost-benefit analysis (in monetary and non-monetary terms). The aim of this approach is to assess and compare different types of CB(RN) risks, select those that are more serious than others, investigate where capability improvements could be made, prioritise measures that could bring about those improvements, and thereby spend a limited budget as effectively as possible. 06/09/2012 : 14:00 : ST.L.Nelson Code: OR54A701 Understanding the Productivity of a Mobile Workforce Dr Jonathan Malpass (BT) Effective organisation of a large mobile workforce requires a vast array of systems, processes and measures. One measure that it is imperative to understand is the productivity of the individuals and teams, identifying reasons for higher levels and recognising how it varies across different geographies. There are various measures of productivity, and this presentation will look at one that has been recently introduced to a UKwide workforce which has led to the need to understand what employees and managers can do in order to affect it. A combination of statistical analysis and modelling tools has been used to provide the business with valuable insight. This presentation will define how productivity is measured, illustrate differences between teams from different areas and provide insight into the different drivers which can be used to improve operations and the business benefits that could be realised. 06/09/2012 : 14:30 : ST.L.Nelson Code: OR54A737 "And the winner is ..." - Case Studies in Tender Evaluation Dr David Wrigley (ORvis Consulting Ltd) Many organisations and public bodies are required to perform competitive procurement to ensure fairness in placing public funds and to ensure good value for money is achieved for the taxpayer. These procurements usually involve some form of tender evaluation scheme which mark the tenders against various criteria and produce an overall score to rank the bidders. The author has used multi-criteria decision analysis (MCDA) to support a wide range of tender evaluations over the past 30 years. This has provided not only the opportunity to see various practices and approaches to decision making in customer organisations, but also to see how O.R. and especially MCDA can support decision making in a range of 98 situations. Several example cases will be presented, and will examine specific challenges and situations including: “ A case where the decision appeared to have already been made; “ A case where the decision appeared to be too close to call; and “ A case where the true value of MCDA proved to be in forestalling a legal challenge. Challenges met in this work have included: over-coming the drawbacks of pre-defined evaluation schemes, having to adjudicate between expert technical evaluators and their managers, and assisting decision makers to understand the limitations of numerical evaluations. MCDA offers a rigorous and auditable approach to tender evaluation which can be used to defend decision makers against disputes, whilst acting as an arbiter of merit which is independent of any biases of the evaluators and procurement managers. 06/09/2012 : 15:00 : ST.L.Nelson Code: OR54A684 Applying Soft Systems Methodology to Better Regulation Mr Ian Mitchell (BIS) The paper offers a case study of O.R. Consultancy. It describes a study by the Operational Research Unit (ORU) within the Strategy, Analysis and Better Regulation Group at the Department of Business Innovation and Skills (BIS), using Peter Checkland’s Soft Systems Methodology (SSM) to support the Better Regulation Executive address strategic questions. The result was a better understanding of the processes and gaps in the creation of regulation and its place as one means in the broader policy context. The paper covers the technical use of SSM, the techniques adopted for this project and considers how social aspects of the work allowed O.R. to assist the management of change. 99 O.R. in Education Organiser: Louise Orpin 04/09/2012 : 11:00 : ST.L. St.Trinneans Code: OR54A575 O.R. in Schools: How to Connect Young People and Maths Dr Vincent Knight (Cardiff University) and Miss Louise Orpin (OR Society) Participants in Science, Technology, Engineering and Mathematics (STEM) subjects need to actively encourage and enthuse young people to explore the various possibilities available to them. Operational Research is no different. In this talk we will present what the OR society already does to raise awareness of O.R. in schools as well as playing a game (audience participation is required) used to introduce game theory to school children. It is hoped that this talk will encourage and enthuse Operational Researchers to participate in such events and further the great work done by the Society to disseminate the subject. 04/09/2012 : 12:00 : ST.L. St.Trinneans Code: OR54A721 O.R. Outreach: British Airways’ Experiences of Promoting O.R. at School, College and University Miss Naomi Crouch (British Airways) Promoting O.R. in schools, colleges and universities is key to ensuring a healthy supply of future O.R. professionals. British Airways recognises this and is committed to raising awareness of O.R. and its applications amongst students, teachers and careers advisors. In this presentation we will share a few examples of the outreach events we participate in. We will also discuss the practicalities involved in doing this (e.g. contacting schools, designing activities, obtaining training) which we hope will be of use to other organisations looking to develop an outreach strategy. 05/09/2012 : 08:30 : JM Salisbury Code: OR54A629 A Statistical Model to Estimate Business and Management Quality Research: Results of Research Assessment Exercise (2001 and 2008) in the UK Prof Alberto Paucar-Caceres (Manchester Metropolitan University) and Dr Xin Shi (Manchester Metropolitan University) We propose a statistical model to estimate the ranking of the Research Exercises submitted to the subject panel of Business & Management Studies by Universities in the UK. We revise and assess the two previous attempts to estimate the journal rankings in business and management by discussing their attempts to ‘replicate’ the panel decisions. We expose the strengths and limitations of these previous attempts carried out by Gaery et al (2004); Kelly et al (2009); and Mingers et al. (2009) who analysed the submissions for the RAE2001 and RAE2008. In this paper, we aim to develop a robust latent model which estimates opinions from the subject panel members and other factors. Expected Maximisation approach will be used to estimate parameters of the statistical model. In addition, the outputs of the model will be compared with different Quality Journal Lists, e.g. Association of Business Schools (ABS) (http://www.theabs.org.uk/?id=257); and Harzing (http://www.harzing.com together with results of the universities rankings in the two previous RAEs to understand the way in which Business & Management Studies might be assessed and ranked in the REF2014. 100 05/09/2012 : 09:00 : JM Salisbury Code: OR54A703 Modelling Latent Preferences when Choice Outcomes Reflect Limited Availability of some Alternatives: A Case Study of School Choice in Northern Ireland Mr Martin Spollen (Queens University Belfast), Dr Karen Cairns (Queens University Belfast) and Prof Adele Marshall (Queens University Belfast) Choice modelling methods reported in the literature rely on the assumption that a decision-maker, i, prefers alternative j within the utility maximising paradigm if i is observed to select j from a given choice set. Where an alternative is unavailable to that decision-maker, this fact is assumed to be known to the researcher ex ante such that the alternative is removed from the individual's choice set prior to modelling. In this sense, the empirical dataset, whether from revealed preference (RP) or stated preference (SP) sources, is assumed to be an unbiased indicator of the underlying preferences in the population under study. The research presented in this paper considers the class of choice-modelling problem where such an assumption cannot be made. These instances occur where the supply of some ('popular') alternatives is inadequate to satisfy aggregate demand such that some decision-makers who prefer a popular alternative must choose a less preferred alternative instead. Data on outcomes from such constrained choice situations are referred to here as Latent Preference (LP) data because the true preferences in the population are masked in the dataset by the artefact of the supply/demand imbalance. LP data is often routinely available from administrative systems and a method to utilise such, while controlling for the artefact, will have benefits over a typical reversion to SP surveys in such instances. A simulation-based strategy is developed to recover the 'true' parameters of a choice model that reflects the unfettered preferences in the population. The approach has the added benefit that it can handle any choice modelling framework whilst also accommodating more complex cases of bilateral (or reciprocal) choice that underlies the rationing mechanism. An application of school choice in Northern Ireland is presented to exemplify the methods developed. This will have wider application to other constrained choice problems. 101 Problem Structuring Methods Organiser: Edward Terry Seagriff 04/09/2012 : 11:00 : JM Salisbury Code: OR54A541 Rich Pictures: Understanding Complexity Through Icons Mrs Tessa Berg (Heriot-Watt University) and Prof Rob Pooley (Heriot-Watt University) Rich pictures are a powerful tool for understanding complexity. The rich picture is not only a tool to aid discussion and debate about problem situations. We claim that understanding the iconography produced during the picturing process leads to a knowledge understanding that far exceeds the limitations of words and speech. We claim there is a language of the rich picture and a subsequent icon grammar structure. 04/09/2012 : 11:30 : JM Salisbury Code: OR54A621 The Valuable Perspectives of Mavericks, Dissidents, and Rebels: A Tutorial on ‚Schools of Thought Analysis‛ (SOTA) Mr Fred Cameron (Centre for Operational Research and Analysis (Canada)) and Dr Geoff Pond (St. Lawrence College) At OR53 the Problem Structuring Methods stream received a demonstration of how Schools of Thought Analysis (SOTA) has been used in the Canadian Department of National Defence. SOTA was used in brainstorming and decision-making activities to tease out when consensus was not as strong as it may have seemed, and to acknowledge and exploit contrarian views to improve decision making. Decision makers need to know when their advice is based on a strong consensus, and when it is not. Many decision-analysis methods give the appearance of consensus when it may be a dysfunctional illusion. Myth-busting methods to shatter such illusions are valuable for improving the decision-making process. This presentation is in the form of a tutorial to give participants the skills to apply SOTA to their own situations. 04/09/2012 : 12:00 : JM Salisbury Code: OR54A752 NATO Guide for Judgement-Based Operational Analysis in Defence Decision Making Prof Neville Curtis (DSTO) Judgement in different guises has been used by military staff whenever assessing problematic situations and making decisions. However, such ‚soft‛ methods may not have had as wide an acceptance within the military community as for instance those based on quantitative and objective techniques. To this end the NATO SAS-087 panel undertook to deliver a ‚Guide‛ to help describe the area of judgement-based analysis; show how the methods can help address problematic situations; and to provide guidance on how a study should be carried out to maximise the validity, credibility and acceptance. A multinational group of senior civilian analysts conducted this activity over a period of two years. The approach taken was to: exploit the ‚puzzles, problems and messes‛ paradigm; develop a model for a ‚soft‛ OA/OR study; identify role and roles and responsibilities of the individuals involved (e.g. analysts and stakeholders); detail the process and provide options for conducting a judgement-based analysis; provide comment on presentation of results and outcomes; and finally to propose the essence of a ‚Code of Best Practice‛. A 140 page analyst-oriented volume is now freely available, summarising this work. In parallel, a 50 page clientoriented volume was also produced. The aim was to provide a familiarisation of the area in terms of three 102 key aspects: that progress can still be made for seemingly intractable and complex decisions; it provides a creative journey of discovery; and how the defence client can accommodate the concerns of scrutiny through a cooperative program with the analyst. The presentation will describe both the study approaches and the detail of the analyst- and client-oriented volumes. The authors believe that these volumes will have widespread applicability to any analysis where there is considerable initial uncertainty, not just the Defence domain. 04/09/2012 : 13:30 : JM Salisbury Code: OR54A579 Getting a Grip on the Problem: Building Student Competence and Confidence in the Application of PSMs Prof Fran Ackermann (Strathclyde University) and Mr James Alexander (Strathclyde University) Teaching soft ‘O.R.’ or Problem Structuring Methods (PSMs) can be a challenge. For example, helping students build both confidence and competence in a method’s use (particularly when there are techniques to master, groups to manage and possibly software to use); enabling students to appreciate the benefit/value added (students typically find the methods lack of ‘a single right answer’ daunting); and contextualising PSM’s within organisational settings, are all important considerations. This presentation discusses how a combination of workbooks, videos, learning check lists and interactive exercises are being developed to alleviate some of the challenges faced by those teaching as well as those studying PSM. 04/09/2012 : 14:00 : JM Salisbury Code: OR54A620 Analysis Based on the Strategic Choice Approach to Evaluate the Operational Effectiveness of the Naval Deployment Model in the Gulf Mr Edward Terry Seagriff (Dstl) Analysis was conducted based on the Strategic Choice Approach to independently confirm that the current naval delivery model in the Gulf is the most effective. The analysis assessed a variety of naval options across a range of factors including objective quantifiable factors (such as costs (both direct and indirect), asset availability, capability effectiveness, personnel issues and training) as well as softer/more subjective assessments (such as social/cultural and wider region engagement). 04/09/2012 : 15:30 : JM Salisbury Code: OR54A518 O.R. for Air Military Assistance to Security and Development (MASD) Dr John Holt (Dstl) The objective of this task has been to focus on air power's contribution to Military Assistance to Security and Development (MASD) operations. The study has included Low-Intensity Conflict (LIC), Counter Insurgency (COIN), and Counter Terrorism (CT), together with stabilisation and security operations. Innovative use has been made of soft O.R., including Cognitive Mapping and Repertory Grid. Cognitive Mapping was used to map concepts and doctrine to form a single unified view. The maps were used to generate questionnaires to interview OA model custodians about their coverage of Air MASD, and for them to provide ratings. The Air MASD factors were also rated in importance using military judgement. From this analysis, eighteen important areas of Air MASD were found to be lacking in model coverage. Clear recommendations have been made on the way forward to provide 'quick wins' for model coverage. 04/09/2012 : 16:00 : JM Salisbury Code: OR54A725 A Decision Making Framework for a Colombian Business Group Dr Julian Benavides Franco (Universidad Icesi) and Dr Juan Felipe Henao (Universidad Icesi) We develop a six step framework to facilitate decision-making. It emerges from an ongoing project with a large business group where data from a survey, focus groups and unstructured interviews was gathered in order to identify the group’s needs in decision-making. The nature of the main problems and decisions the firm faces were identified as: 1) Generation of strategic initiatives; 2) Assessment of internal and external 103 issues to identify SWOT; and 3) Evaluation of strategic initiatives. Thus, an initial framework was formulated; it combines approaches of strategic management, Soft O.R. and decision analysis. Next, three major corporate decisions were successfully undertaken following the proposed guidelines: 1) A prioritisation of a portfolio of projects, over a five year horizon with multiple project types and financial and non-financial criteria; 2) Main technological supplier selection for a new venture with multiple criteria and cash flow uncertainty; and 3) Disposal of a failed initiative, considering several stakeholders, alternatives generation and uncertainty. Following the implementation results, the framework was fine tuned. The resulting six final steps are: problem formulation and definition, strategic goal vs. decision objectives coherence, SWOT and stakeholder analysis, alternatives generation, alternatives evaluation and scenario analysis. The method is not necessarily linear, and steps 1 through 4 add to define relevant criteria. Alternatives are ranked using the benefit to cost ratio and strength tested under scenario planning and risk, regret, and robustness criteria. Additional actions are being taken to ensure the transferability of the framework and its application across the entire organisation. 04/09/2012 : 16:30 : JM Salisbury Code: OR54A736 Empirical Simulation Methods to Identify Issues in a Two-Sided Military System Prof Neville Curtis (DSTO) When faced with a new study into a military system of concern the critical first step is to identify the ‘issues’ involved and to develop models of the system. We have used two empirical methods for this in the past: one based on a sequential examination of the stages involved in conducting military actions, and the second focussed on examining the functional higher-level contributions to achieve military outcomes described in an influence diagram. Both methods are compatible with the notion of a two-sided interaction between blue and red forces and rely upon subjective and speculative interpretation of an incompletely understood system. In this paper we illustrate the strengths and weaknesses of the two methods through the example of military operations conducted in an improvised explosives devices (IEDs) environment. While both methods lead to discovery of similar issues, they emphasise different aspects with, for instance, the sequential method placing more weight on ‘kinetic’ effects, goes into depth more than breadth, may be more useful to identify immediate deficiencies and finally may be more amenable as a discussion item with the stakeholders. In contrast, the influence diagram approach leads to a different balance of concerns, is more adaptable to future perturbations and has more potential to lead to semi-quantitative insights on priorities. We believe that the two methods are more suited to the ‘requirements’ and ‘needs’ (conceptual) phases, respectively. Such empirical methods will have practical application to similar studies at the ‘frontend’, including those involving non-military themes. 104 Project Management Gary Bell Rosanne Pagano Jon Warwick Organisers: Gary Bell, Rosane Pagano and Jonathan Warwick 04/09/2012 : 11:30 : ST.L.Bonnar Code: OR54A528 Cultural Diversity and Workplace Dynamics: The Case of Transmanche Link Dr James Freeman (Manchester Business School) The paper addresses workplace dynamics in a multi-cultural establishment. Specifically, data collected from the British-French construction consortium, TransManche Link is analysed to gain insights into the management setting, work climate, employee perceptions and their inter-relationships. Further, differences between the British and French managers on these dimensions are investigated. In order to tackle the complexity of the analysis, the study employed structural equation modelling to hypothesise and simultaneously model the multiple relationships of concern. 04/09/2012 : 12:00 : ST.L.Bonnar Code: OR54A668 An Intellectual Framework For Critically Appreciating Software Project Management Methodologies Dr Gary Bell (London South Bank University) and Prof Jon Warwick (London South Bank University) This paper offers an Intellectual Framework as a broad approach to benchmarking software project management methodologies. Software projects have been associated with the symptoms (or effects) of cost overruns, schedule slippage and poor quality. There are various possible reasons (or causes) of these symptoms, e.g. poor understanding of management methodology. We have decided to benchmark the Goal/Question/Metric (GQM) Paradigm using the Intellectual Framework. This improvement methodology was developed at NASA over a period of 25 years. Victor Basili and co-researchers have published numerous articles of their findings and important evolutions of this methodology. The GQM Paradigm is still being applied at the Fraunhofer Institute for Experimental Software Engineering. However, we focus upon the collated NASA work. The Intellectual Framework has three distinctive parts, namely: a paradigm map, a set of management metaphors and an intellectual device. The paradigm map highlights different social theories and underpinning philosophical assumptions. Moreover, social theories can be linked with respective management metaphors. Management metaphors are ‘a way of seeing’ and ‘not seeing’, which is a very important idea that needs to be explored. The offered intellectual device guides the studying of a methodology in practice. The original device had three key parts, namely: Framework of Ideas, Methodology and Area of Concern are discussed. Three more parts are proposed labelled Organised Learning, Cost Benefit Appraisal and Knowledge Library and their importance are substantiated. Moreover, the device is underwritten by the single and double loop learning concepts. The appreciation aspect of the benchmarking establishes the scientific philosophical roots, connectivity with social theory and appropriate management metaphor of the GQM Paradigm. We also uncover the key concept and principles of this improvement methodology. Additionally, the research strategy, linked with the GQM Paradigm, is identified which ensures rigour in the results. The critical aspect examines the benefit and cosy of implementing this improvement approach. Our benchmarking findings are presented, which can inform changes to this methodology. Additionally, we reflect upon the use of the Intellectual Framework, and its utilisation within the Project Management Discipline. 105 04/09/2012 : 13:30 : ST.L.Bonnar Code: OR54A618 Project Success: An Ethical Perspective Dr Rosane Pagano (Manchester Metropolitan University Business School) and Dr I. K. White (Hull University) The recognition of the ‘iron triangle’ (time, cost, quality) as the main criteria against which to measure project success has been the subject of debate for some time in academic and professional spheres. The importance of stakeholder management as a critical success factor has increasingly come to the fore, approaching in relevance that traditional triangle. This paper focuses on projects whose outcome is ‘change’, rather than a tangible product, and argues that change projects in particular call for the enactment of an ethical professional pledge on the part of the project manager, beyond the skill level of handling people. Change projects often include a subset of stakeholders who will be themselves agents of change or the subjects of a transformative act. For managing those stakeholders, we argue that the project manager’s success relies on transcending ” through the professional pledge ” its role as purely an expert in monitoring and controlling time, cost and quality. In this paper a stakeholder management framework is proposed, based on theories of collaborative learning, to support the implementation in practice of a professional ethic. 04/09/2012 : 14:00 : ST.L.Bonnar Code: OR54A749 A Methodology for Simulation Conceptual Modelling that Embeds the SCOR Process Reference Model Dr Miles Weaver (Edinburgh Napier University), Dr Pavel Albores (Aston Business School) and Dr Doug Love (Aston University) A seven phase methodology is presented for simulation conceptual modelling that combines a prescribed procedure with domain-specific knowledge embedded in the form of the Supply Chain Operations Reference (SCOR) model. The paper builds upon previous contributions that detail the design and development of the methodology using an iterative triangulation method (See Lewis, 1998). This includes incorporating existing simulation conceptual modelling practice (e.g. principles, methods of simplification, frameworks and validation methods); requirements for an ‘effective’ conceptual model; a ‘good’ methodology and the utilisation of a process reference model to improve the efficiency and focus during the process of conceptual modelling. This paper presents an illustration of the methodology with a complex and detailed SCM application. The implication for further refinement and validation of the methodology (i.e. its feasibility and utility) with a range of process reference models in different industrial contexts is outlined. 04/09/2012 : 14:30 : ST.L.Bonnar Code: OR54A707 Exploring The Connectivity Between Holon Rich Pictures, Project Product Descriptions and Product Breakdown Structures Dr Gary Bell (London South Bank University) and Prof Jon Warwick (London South Bank University) This presentation discusses the connection between Holon Rich Pictures, Project Product Descriptions and Product Breakdown Structures through a higher education project. An Intellectual Framework is employed to guide this investigation. Furthermore, the use of different methods for scoping a project can be linked with Methodological Pluralism. Holon Rich Pictures are underpinned by the ‘value-full’ notion ” for they aim to capture a broad understanding of a problematic situation (‘the whats’) from different stakeholder perspectives such users and suppliers. PRINCE2 (Project IN Controlled Environment) is an established project management approach that has seven key principles and themes. One principle is ‘focus on the product’ which has links with the Quality (‘the what’) and Planning (‘the where’, ‘the how’, and ‘the who’) theme. The Product-based Planning method guides the generation of the Project Product Description, Product Breakdown Structures, Requirements and Specifications. It is believed that the use of the Holon Rich Pictures will inform the development of these specialist’s PRINCE2 management products. The mandate (‘the why’) for the higher education project guides the Holon Rich Picture method. The findings generated by the Holon Rich Pictures are outlined which informs the Project Product Description and Product 106 Breakdown Structure. An evaluation of this research is undertaken, which engenders future developmental work. 04/09/2012 : 15:30 : ST.L.Bonnar Code: OR54A715 KEYNOTE: Lessons for Project Management: Experiences from Forensic Modelling Prof Colin Eden (University of Strathclyde) and Prof Susan Howick (University of Strathclyde) Time and cost overruns on complex projects continue to be a challenge for project managers ” the Edinburgh Tram system provides just one recent example. Although complex projects can be difficult to unravel, system dynamics modelling has a history of being able to explain the key events that trigger overruns and their impact. This presentation will demonstrate that the process of building a system dynamics modelling can provide powerful lessons regarding project overruns when the model is built with a focus on ‘telling the story’ of the project. The process of building a model to tell the ‘story’ of the project highlights a number of issues that are often not considered due to the difficulty in addressing them. Addressing these issues can draw out important lessons for the organisation and indicate ways of better managing the planning stage of the project as well as after disruptions have occurred. For example, ensuring that the model; can start telling the ‘story’ of the project by commencing with a validation of the estimate; replicates the decision making process as it occurred in practice, and can replicate memorable events during a project through use of time based scenarios. Using the extensive experience of the presenters in developing system dynamics models for litigation, on behalf of contractors and customers, the presentation will discuss each of the above issues and demonstrate the types of organisational lessons that can be gained during the modelling process. 05/09/2012 : 11:00 : ST.L.St. Trinneans Code: OR54A650 Planning with the OI Factor Ms Michelle Morris (Manchester Metropolitan University Business School) Perception management and intuition have been highlighted in the literature as critical abilities for managers and potentially critical to project stakeholder management. This influence extends optimisation methods in planning and presents the possibility of adding ‘organisational intuition’ (OI) ” intuition related to a specific organisation or industry sector ” to the planning process for all that organisation’s projects. Planning is the most important part of project management, the development of the plan takes time and effort (usually between 5-10% of a project). To what extent can it be shortcut by using OI from one project plan to another? When there is a growing recognition of the influence of human factors on keeping projects to plan, transferability from past projects to new ones is desirable, but is it achievable? There is a case for capturing organisation intuition, which could be applied to the planning of projects for that organisation or industry sector. The current challenge is to formalise a template and create a framework, a hybrid of existing methodologies. In this paper a number of social theory techniques together with a number of emerging ethnographic methods have been evaluated and the beginning of a framework is starting to emerge. 107 Queue Modelling Organiser: Navid Izady 05/09/2012 : 08:30 : ST.L.Brewster Code: OR54A631 Co-ordinated Operation of Queues at Congested Arterial Signalised Intersections Mr César Dario Velandia (The University of Nottingham), Dr Jason Atkin (University of Nottingham), Dr Ruibin Bai (University of Nottingham) and Prof Graham Kendall (University of Nottingham) Urban traffic congestion is one the most pressing issues in terms of transportation in cities. Traffic control is essential to tackle this problem and an optimised traffic signal timing strategy is one of the most costeffective ways to achieve improvements. The importance of queue management strategies at signalised road intersections is well-known in this context and modern traffic control strategies are capable of handling queueing vehicles efficiently during light traffic conditions. The most common approach to intersection management is to minimise the overall delays, primarily considering the aggregated delay of vehicles waiting during the red phase for each approach. These strategies are usually limited to isolated intersections, but attention to network-wide solution has steadily increased. In particular, arterial road management is challenging due to the number of movements (lane groups) to co-ordinate at each junction and the inherent variations in arrival rates and trajectories (e.g. straight on, right turners, etc.). To overcome the limitations of other queueing models at arterial roads, it is necessary to devise control strategies which account for prevailing intersections conditions and moving queues (platoons) of vehicles approaching from upstream. Furthermore, adaptive management uses multiple queue control objectives, the most commonly used being to minimise delay, maximise capacity or synchronise queue clearance. Effective signal coordination of congested arterial traffic involves precise management of vehicular queues. For this reason, we propose a model to calculate offsets (the time difference between the start times of cycles of consecutive co-ordinated signals) in arterial facilities so that (i) weighted objectives based on the prevalent queue conditions are minimised, and (ii) smooth (unstopped) arterial movements are attained, by synchronising turning movements and overflown queues. This approach will be studied in detail using micro-simulation and a wide range of scenarios and intersection conditions. 05/09/2012 : 09:00 : ST.L.Brewster Code: OR54A706 When to Use Virtual Hold Technology in Call Centre Operations Dr David Worthington (Lancaster University) and Dr Chris Kirkbride (Lancaster University) At busy times call centres can be temporarily overloaded, leading to substantial queues building up and callers abandoning their calls. If acquisition calls abandon this may well lead to a loss of sales. If care calls abandon, customers may well be disappointed by their experience, and in the longer term may look for alternative suppliers and pass their experience on to other potential customers. Virtual Hold Technology (VHT) offers the customer the option of being rung back without losing their place in the queue. If they do not select this option, they can still either wait in the queue or abandon their call. The aim of VHT is to reduce abandoned call rates and increase customer satisfaction. However, it is possible that conditions might arise, typically with higher call volumes and larger numbers of customers in VHT, where overall call wait times (ASA) might significantly increase as a result of lower abandon rates associated with VHT customers, leading to increased dissatisfaction amongst all customers. This paper describes queue modelling work of various types to attempt to predict the performance of the call centre in terms of 108 abandon rates and ASA’s for scenarios such as: 1. VHT enabled on all queues; 2. VHT disabled on all queues; and 3. VHT enabled on acquisitions only. 06/09/2012 : 08:30 : ST.L.Bonnar Code: OR54A594 Price of Anarchy in Queueing Systems Mr Rob Shone (Cardiff University) A queueing system can be represented as a stochastic process, in which events (e.g. arrivals, service completions) occur at random points in time. Problems related to the optimal control of queueing systems are usually concerned with the optimisation of an overall summary measure of the system’s performance, such as the long-run average cost incurred per unit time. These types of problem can be modelled using Markov Decision Processes (MDPs), in which we seek to determine an optimal "policy" which specifies the most advantageous decisions to be taken by the system controller. Decisions may involve the admission (or denial) of service to individual customers, the routing of customers between different service facilities and the scheduling of service priorities. In this talk we briefly discuss methods of finding optimal policies for queueing systems, and some of the structural properties of these optimal solutions. We also discuss the general principle that the performance of a queueing system as a whole is not optimised if customers are allowed to make "selfish" decisions which maximise their own expected net gain. A measure of the suboptimality of selfish (myopic) policies is given by the 'Price of Anarchy' (PoA), which is simply a ratio of the long-run average cost under a 'selfish' policy to the average cost under an optimal policy. The concept of PoA has been studied in various game theoretical settings, and in this talk we discuss its applicability to queueing systems and MDPs in general. 06/09/2012 : 09:00 : ST.L.Bonnar Code: OR54A642 Understanding the Effect of Individual Behaviour in Hierarchical Queues Mr Jason Young (Cardiff University) and Dr Vincent Knight (Cardiff University) A variety of real world services are in fact a network of hierarchical queues where players (customers) can choose to skip certain levels of the process at a particular cost to their utility. An example of this is the boarding process for budget airlines as well as certain healthcare processes. It is well understood that selfish behaviour can hinder the overall performance of queueing systems. In the simplest of queueing situation where players can choose whether or not to join a queue selfish behaviour can have a detrimental effect. The subject of this talk will be to measure the inefficiency caused by selfish individuals in hierarchical queueing systems where the strategies available to each player is not only whether or not to join the queue but also at which level to join. The approach presented is a simulation approach but other methodologies will also be discussed including Markov decision processes as well as routing games. 06/09/2012 : 09:30 : ST.L.Bonnar Code: OR54A698 A Queueing Model for Approximating Indirect Waiting Times in Specialty Clinics Dr Navid Izady (University of Southampton) In this presentation we propose a discrete bulk service queueing model for approximating the indirect waiting time distribution of elective patients in specialty clinics. The indirect waiting time is defined as the time between referral of an elective patient by a primary care physician and his/her first visit to the clinic. The model captures the behaviour of no-shows by assuming that every patient, independently of other patients, does not show up with a fixed probability. To measure the waiting time of those patients that do not turn up for their appointments and are subsequently re-scheduled, in line with NHS guidelines we assume the referral date is the date of their last missed appointment. Using data from a specialty clinic, we show how the model could help in deciding on the number of patients of various types that should be seen in every clinic session. 109 Routing Applications & Transportation Luc Muyldermans Organisers: Luc Muyldermans and Jamal Ouenniche 04/09/2012 : 11:00 : ST.L.Cowan Code: OR54A529 Design of Hybrid Metaheuristics for Vehicle Routing Problems Mr Prasanna Kumar Ramaswamy (University of Edinburgh Business School) and Dr Jamal Ouenniche (University of Edinburgh Business School) Several exact algorithms were designed to solve routing and scheduling problems such as branch-andbound and branch-and-cut algorithms. Given the combinatorial nature of these problems, exact algorithms are typically used to solve relatively small to medium sized instances, whereas construction and improvement heuristics such as meta-heuristics are typically used to solve large problems. So far, research on meta-heuristics has shifted from the design of ‚pure‛ meta-heuristics (e.g., simulated annealing, tabu search, variable neighbourhood search, genetic algorithms) to the design of hybrids where one integrates within the same search framework several search strategies of the pure meta-heuristics so as to exploit the strengths of various methods or search strategies embedded within the hybrid design. In this study we develop a new hybrid meta-heuristic framework and test its performance in solving vehicle routing problems (VRPs). To be more specific, we explore different ways of designing hybrid meta-heuristics. The main question we intend to answer is as follows: Is it possible to use the intelligence of exact algorithms to guide the search for an optimal or near optimal solution to VRPs? Early computational results seem to confirm the attractiveness of this type of hybrid design in addressing Routing and Scheduling problems. 04/09/2012 : 11:30 : ST.L.Cowan Code: OR54A599 Dynamic Congestion Charging: Investigating Boundary Effects when Applying Low-revenue Tolls Dr Kathryn Stewart (Edinburgh Napier University) The classical road tolling problem in economics is to toll network links such that under the principles of Wardropian User Equilibrium assignment a System Optimising flow pattern is obtained. In the static case, the minimal revenue toll problem is capable of solution by various methods, such as linear programming and heuristically by reduction to a multi-commodity max-flow problem. Previous work by Stewart (Stewart, 2007) in a static stochastic environment has developed heuristics to determine low revenue tolls that produce good sub-optimal flow patterns where total network costs approach the system optimal. Dynamic Traffic Assignment models extend the static concept of UE to DUE (dynamic user equilibrium) where the system is said to be in DUE where no user can unilaterally reduce their origin to destination travel cost. Algorithms for DUE commonly iterate between two components; the Dynamic network loading and route choice or path inflow reassignment. This paper utilises existing DUE algorithms (Carey and Ge, 2011) to investigate their ability to incorporate tolling scenarios and to produce low-revenue tolls to create desired flow profiles using network models. This paper investigates tolling strategies under different assumptions and determines toll sets which reduce total network cost. Toll profiles that are both flat and bell-shaped are examined with reference to their impact on undesired spatial and temporal boundary effects. Scenarios are examined where firstly tolls must be levied across the entire time horizon, and secondly where the tolling 110 period is a sub-period of the modelled time. Two tolling regimes are examined; where all paths may be charged and when a single path must remain toll-free. 04/09/2012 : 12:00 : ST.L.Cowan Code: OR54A638 A Meta-Heuristic to Solve Vehicle Routing Problems with Dependent Vehicles Mr Edward Kent (The University of Nottingham) and Dr Jason Atkin (University Of Nottingham) Most studies of the Vehicle Routing Problem with Time Windows (VRPTW) consider independent vehicles making independent visits to depots and customers. These problems have been solved in many different ways in the past using heuristics and exact algorithms. In this talk we present a real-world variant of the VRPTW where there are both dependencies between vehicles/visits and product mixing rules. Customers and depots which only have a limited amount of loading bay space are restricted in how many vehicles they can have on site at once. Furthermore, certain products cannot be carried in the same vehicle at the same time for reasons of contamination risks or safety. Both of these elements introduce dependencies between vehicles in order to avoid waiting for a loading bay to be free. On a busy morning, capacity limitations at a depot may mean that some vehicles have to leave early in order to prevent others leaving too late to achieve all of their time windows. There may also be a trade-off required between different customers if capacity limits mean that not all deliveries can be made. The problem can be presented mathematically as a 3-index model (heterogeneous) VRPTW with extra constraints to handle the mixing rules and loading bay rules. The problems were found to be too hard for IP solvers such as CPlex. We present a generic framework for modelling these and other vehicle routing problems, along with a meta-heuristic solution method, comparing it against some greedy solution approaches. 04/09/2012 : 15:30 : ST.L.Cowan Code: OR54A572 School Bus Routing Case Study using Ant Colony Optimisation Mr Jose Jimenez (Pontificia Universidad Javeriana), Mr Jairo R Montoya-Torres (Universidad de la Sabana) and Mr Juan Sebastian Arias-Rojas (Pontificia Universidad Javeriana) This practical case study aims to schedule efficiently a fleet of buses of certain school in Bogota (Colombia), as a school bus routing problem (SBRP). Even is considered as two separated problems, this combinatorial optimisation case study includes the early pick up students from various bus stops and the delivery of them afterwards at afternoon. It considers the following constraints: maximum capacity of the bus, maximum riding time of students and time window to arrive to school/student home; among the regular Vehicle Routing Problem. The problem described is solved using Ant Colony Optimisation (ACO). It is solved using a two-phase resolution approach. The first phase consists on define the assignment of student pickup (or student delivery) points to buses, while the second phase consist on the actual routing of buses using an ACO based algorithm. During the resolution, a sensitivity analysis is carried out in order to validate the parameters chosen to run the algorithm. All the computational experiments are performed using real data. Results leads to increased bus utilisation and reduction in transportation times with on-time delivery to the school. The proposed decision-aid tool has shown its usefulness for actual decision-making at the school: it outperforms current routing by reducing the total distance travelled by 8.3% and 21.4% respectively in the morning and in the afternoon. 04/09/2012 : 16:00 : ST.L.Cowan Code: OR54A628 Variants of the Capacitated Arc Routing Problem: An Overview Dr Gu Pang (Newcastle University) and Dr Luc Muyldermans (University of Nottingham) The Capacitate Arc Routing Problem (CARP), originally proposed by Golden and Wong (1981), is a wellknown combinatorial problem in which demands are associated with the edges of a graph, vehicles have limited capacity and the goal is to construct routes of total minimum distance, each starting and ending at the depot and such that the total demand serviced by a route does not exceed the vehicle capacity. Over the past decades several variants and extensions of the standard CARP have been proposed in the literature, essentially, to model more realistic real life situations. These variants include: the stochastic CARP, the 111 CARP with time windows, the CARP with time-dependant service costs, the periodic CARP, the bi-objective CARP, the CARP with intermediate facilities, the split delivery CARP and the multi-compartment CARP. The purpose of this presentation is to provide an overview of these CARP variants, together with a flavour of the solution techniques and algorithms that have been proposed to solve these problems. 04/09/2012 : 16:30 : ST.L.Cowan Code: OR54A632 Periodical Vehicle Routing Problem Due to Driver Familiarity Mr Matthew Soulby (University of Nottingham) and Dr Jason Atkin (University of Nottingham) Information obtained from the vehicle routing software and consultancy company, Optrak, has allowed us to generate a more accurate, rich representation of a Vehicle Routing Problem (VRP) which is encountered by delivery companies. Driver familiarity is often of great importance to delivery companies. Not only will familiarity decrease service and travel times, due to knowing the roads, building entrances, delivery points and so on, but the rapport which can be built up between driver and customer also has value. When driver learning is applied to the VRP a trade-off occurs between route length and driver knowledge about the route. For the normal VRP, where driver familiarity is ignored, there are no benefits from maintaining the same or similar routes, so taking account of the periodical nature and more predictable elements of the stochastic customer demand is of little consequence. To be able to determine the effect that driver familiarity has upon solution quality, it is not possible to consider only a static problem. Familiarity changes over time, and assessment methods must take this into account. Our method considers solutions in terms of any similarities that occur across days, such as the allocation of drivers and vehicles to given routes, the combination and order of customers within each route and the consequent estimated arrival times at customers. To demonstrate this method, an evolutionary algorithm has been applied to the problem, with and without simulating the benefits of driver familiarity, within a generic VRP framework which we developed. We have considered the application of the algorithm with and without an objective to improve familiarity. We will discuss the ways in which the addition of these familiarity objectives affects the solutions, and the potential benefits of aiming to use familiar drivers and improve driver familiarity when solving such real world problems. 06/09/2012 : 12:00 : ST.L.Bonnar Code: OR54A647 Lower and Upper Bounds for the Joint Problem of Request Routing and Client Assignment in Content Distribution Networks Mrs Narges Haghi (University of Southampton) A Content Distribution Network is a system of servers containing digital objects that are placed at selected nodes of a telecommunications network. This paper describes a nonlinear integer programming model to solve the joint problem of request routing and client assignment which explicitly considers delays in transmitting objects. The paper also describes a lower bounding procedure for the problem based on Lagrangian relaxation and subgradient optimisation, as well as an Iterated Local Search algorithm to generate upper bounds. Results of computational experimentation obtained by using randomly generated data will be presented at the conference. 06/09/2012 : 13:30 : ST.L.Bonnar Code: OR54A635 Receding Horizon Approach to Gate Allocation Problem Mrs Urszula Neuman (University of Nottingham), Dr Jason Atkin (University of Nottingham) and Prof Edmund Burke (University of Stirling) The gate allocation problem is one of the more important airport operation problems and has had considerable research undertaken in the past. It involves finding appropriate gates/stands for aircraft arriving at an airport. The gate allocation systems which are currently used at many airports do not support the real-time online decision process that controllers face on the day of operation when they often have to change the previously prepared plan to recover from delays or other changes. Hence, this research focuses on designing and evaluating a fast search methodology that finds a good solution quality. We initially 112 consider the case where there are fewer side constraints and preferences, resulting in a problem where they cannot be used to prune the search space. The investigated method is instead based upon a receding horizon approach, considering a time-window-based decomposition of the problem. This method corresponds well with the nature of the real world problem, since knowledge about future situations, may be limited as some aspects of the gate allocation problem (e.g. delays) are only known about very shortly in advance. It performs well despite this limitation. The most important parameter of the receding horizon approach is the distance to the horizon, called the time window size. This size influences both the calculation time and the quality of the solution. Its effects are investigated in depth in this research. A real data set containing arrival and departure details for one terminal has been used in this investigation. The problems are modelled as mixed integer programming models and solved using CPLEX solver. The formulation is modified during the solution process to ensure an acceptable solution speed. We will discuss this approach and compare the results obtained using the receding horizon approach against optimal solutions obtained by solving the whole problem at once. 06/09/2012 : 14:00 : ST.L.Bonnar Code: OR54A689 Methods for Solving the Dynamic Vehicle Routing Problem with Pickups, Deliveries and Time Windows Miss Penny Holborn (Cardiff University), Dr Rhyd Lewis (Cardiff University) and Dr Jonathan Thompson (Cardiff University) Due to the increasing demands on transportation services to deliver a fast and efficient service, systems are needed for dispatching transportation requests that arrive dynamically throughout the planning period. The focus of our research is the Dynamic Pickup and Delivery problem with Time Windows (DPDPTW), where requests are not completely known in advance but become available during the planning period. All requests have to be satisfied by a given fleet of vehicles and each request has a pickup and delivery location, along with a time window at which service can take place. To solve the DPDPTW we are investigating methods embedded in a rolling horizon framework, thus allowing us to view the problem as a series of static ones. Initial research concentrated on the static variant of the problem where both exact and heuristics methods were applied to solve the problem. Our current research is dedicated to solving the dynamic problem, where a time stamp is allocated to each request and the request does not become known to the system until that time. Investigations have been performed to identify when the algorithms should be updated to incorporate the arrival of new requests and how this request should then be incorporated into the existing schedule. Datasets with both varying degrees of urgency and proportion of dynamic requests have been examined along with various exact and heuristic methods. Competitive results have been achieved across a range of benchmark datasets. This continually increasing area of research covers many real life problems such as the Health Care service, with the collection and delivery of patients, specimens and equipment between hospitals and Health Centres. Our current research also looks at applying our methods to solve a real life problem for a local Health Courier service. 06/09/2012 : 14:30 : ST.L.Bonnar Code: OR54A735 Heuristic Route Generator for the Workforce Scheduling and Routing Problem Mr J. Arturo Castillo-Salazar (University of Nottingham) and Dr Dario Landa-Silva (The University of Nottingham) In the context of workforce scheduling and routing problems (WSRP), those which require a mobile workforce to perform skilled activities at different locations and with a specific time window. Various solution methods have been used to approach both the routing component and the employee scheduling one. This abstract presents an algorithm that generates routes for employees. All routes are feasible since they are created for individual employees. The routes take into account some of the different types of constraints that can be found in WSRP. Among the constraints tackled are: enforced time windows, skillrelated, preferences and domain dependant constraints. Two evaluations criteria for routes are presented, the first one based on cost of the route and the second one based on the violation of preferences of the recipient of the activity. The algorithm has two stages: initial and improvement. 113 Scheduling Organisers: Djamila Ouelhadj 04/09/2012 : 13:30 : JM Duddingston Code: OR54A559 Bi-criteria Forge Scheduling of Jobs with Tabu Search Dr Rasaratnam Logendran (Oregon State University) and Miss Yasaman Mehravaran (Oregon State University) This research is motivated by a real industry problem on forge scheduling. The goal is to find the best sequence and assignment of jobs to machines in order to minimise both producer's work in process inventory and maximise customers’ service level, thus resulting in a bi-criteria objective function. This problem has a challenging structure with eight different stages. A job passes through stage one and then goes to stage two. After finishing stage two, a job can revisit stage one and again go through stage two. This revisiting can occur several times depending upon the type of job. The job then proceeds to stages three and four. After finishing stage four, a job can cycle back to stage one and restart its operations just like the first cycle. The cycling back can only happen once. Then the job goes through stages five and six. After finishing stage six, a job can revisit stages five and six for several times. Finally, the job passes through stages seven and eight. Stages one, three and five have unrelated-parallel machines, while there is only one machine in the other stages. The setup time is sequence-dependent in some stages. Also jobs are allowed to skip stages. The job release and machine availability times are considered to be dynamic. Depending on job’s type, a job can be split into a maximum of six pieces when it passes through one or more of its revisits to stages one and two. Also the machines in stages one and five are capable of processing at most six split pieces at the same time. In this research we develop a search algorithm based on tabu search to find the optimal/near optimal schedule. Random problem instances are generated from the real industry data to assess the effectiveness of the search algorithm. 04/09/2012 : 14:00 : JM Duddingston Code: OR54A660 Scheduling RAF Air Traffic Control Personnel Mr Richard Conniss (University of Nottingham), Dr Tim Curtois (University of Nottingham) and Prof Sanja Petrovic (University of Nottingham) Research in personnel scheduling has largely focussed on specific industries or professions. The available literature includes many examples of work done on nurse rostering and aircrew scheduling. An interesting addition to this area is the rostering of military Air Traffic Control (ATC) personnel, mainly due to the complicated nature of the set of constraints and goals to be met. Rostering ATC personnel has some similarities with both nurse and aircrew scheduling; all are subject to shift patterns, qualifications, working hours restrictions and rest break planning. The differences make this a unique problem, with some novel, challenging and interesting issues that require further exploration. Each ATC unit is made up of a set of control positions and controllers. Each controller holds a set of qualifications that allow them to work in one or more positions to achieve a specific ATC task. To obtain a qualification, a controller is designated as under training (UT) in a position, and then must be supervised by another controller who holds a qualification for that position and an additional unit instructor (UI) rating in that position. UI ratings are specific to a particular position, not all controllers will hold UI ratings in all of their qualified positions. Also, the UT may well hold qualifications in other positions and whilst under training cannot be used by the unit 114 to staff other positions. The aim of our research is to produce a system to manage the scheduling of ATC personnel. This will require shift planning, but also scheduling which position an individual will occupy, providing adequate rest breaks for controllers whilst maintaining operations, and forecasting training to maintain sufficient qualifications for ATC operations. We will present a detailed model for the problem, early results and discuss possible areas for further research. 04/09/2012 : 14:30 : JM Duddingston Code: OR54A591 A Genetic Algorithm for Maintenance Scheduling of Generating Units Dr Khaled Alhamad (PAAET - Kuwait) The load of the power system increases rapidly, along with the rapid development of the industry. Cogeneration for power generation and fresh water production has been used by utility companies for many years as a reliable, efficient and economic means of generating power and simultaneously desalting sea water. An important problem in cogeneration plant is the long term Maintenance Scheduling (MS) of the generation and desalination units. The generator maintenance scheduling problem is to determine the period for which units should be taken off line for planned maintenance over the course of one year planning horizon. In order to minimise the time and increase the system reliability, system load demand constraints must be satisfied. In the recent decade, many efforts have been done in the maintenance scheduling field. Alardhi and Hannam proposed methodology based on mixed integer programming model which finds the maximum number of available power and desalting units. To test the proposed model, an example is used to compare two Kuwaiti cogeneration plants for total 42 units. Fetanat and Shafipour presented a formulation that enables Ant Colony Optimisation for Continuous Domains (ACOR) to seek the optimal solution of the unit maintenance scheduling problem. The objective function of this algorithm considers the effect of economy as well as reliability. Various constraints such as spinning reserve, duration of maintenance crew are being taken into account. The objective of this research is to maximise the available number of operational units in each plant. This problem is solved using Genetic Algorithm (GA) heuristic method and a good results have been reached. The proposed model was applied on Kuwait Electricity and Water ministry. 04/09/2012 : 15:30 : JM Duddingston Code: OR54A533 A Local Search Heuristic to Compute Approximate Nash Equilibrium of Competitive Travelling Salesmen Problem Dr Jiawei Li (University of Nottingham) In a competitive travelling salesmen problem, multiple self-interested travelling salesmen compete with each other in visiting a number of cities. The salesmen receive a benefit if they are the first one to visit a city and they will pay a cost for the distance travelled. The objective of each salesman is to gain a benefit from as many cities as possible with the minimum traveling distance. Due to the conflict of interest among multiple agents, the competitive travelling salesmen problem (CTSP) is a dynamic non-cooperative game and the solution is a Nash equilibrium. In this paper, the existence of a pure strategy Nash equilibrium in CTSP is deduced. Due to the complexity in computing Nash equilibrium, a local search heuristic is developed to compute an approximate Nash equilibrium of CTSPs. The approach starts with a non-equilibrium solution that is constructed by assuming that all agents adopt a nearest neighbour heuristic, and then repeatedly changes one move of the solution to another city in the neighbourhood to find a better solution. Several computational examples are given to illustrate the effectiveness of the proposed approach. 04/09/2012 : 16:00 : JM Duddingston Code: OR54A567 A Lookahead Heuristic for the Minimisation of Open Stacks Problem Prof Marco Carvalho (Federal University of Ouro Preto) and Prof Nei Soma (Aeronautic Institute of Technology) We present a new heuristic for the Minimisation of Open Stacks Problem (MOSP), that appears in settings of sequencing patterns and it is related to cutting and packing problems. The proposed heuristic generates 115 an initial solution by applying a breadth-first search on specific graphs ” a new approach for the MOSP ” and then uses that solution by looking ahead to postpone the opening of new stacks and to move up the closure of the already open ones. The suggested heuristic is compared with the winner of the First Modelling Challenge on the MOSP (2005) that is an exact method and with another one that is considered the best performing heuristic from the literature. Two sets of benchmark instances from the literature were used and the results show that the gap between the proposed method and the optimal solution for the first set that has almost six thousand instances is just 0.18%. Moreover, in 97.21% of the cases the optimal solution was found within a maximum 2 seconds of running time and to those cases where the optimal solution was not found the differences were no more than 2 stacks. For the second set of 200 instances, the gap between the best performing heuristic from the literature and the proposed heuristic is 6%, with a maximum running time of 33 seconds. 04/09/2012 : 16:30 : JM Duddingston Code: OR54A661 Solving Stochastic Unit Commitment by Column Generation Mr Tim Schulze (The University of Edinburgh), Dr Andreas Grothey (The University of Edinburgh) and Prof Ken McKinnon (The University of Edinburgh) In recent years the expansion of energy supplies from volatile renewable sources has triggered an increased interest in stochastic optimisation models for generation unit commitment. Several studies have modelled the problem as a stochastic mixed-integer (piecewise linear or convex quadratic) multistage problem. Solving this problem directly is computationally intractable for large instances and many alternative approaches have been proposed. However, few of them exploit the structure of the multistage formulation. In this talk we outline how a Dantzig-Wolfe reformulation can be used to decompose the stochastic problem by scenarios. We develop a Column Generation framework which can handle quadratic generation costs and is capable of solving stochastic unit commitment problems to optimality. Numerical results are given to illustrate that convergence can be achieved within a few iterations of our method. 116 Simulation Organiser: Kathryn Hoad 04/09/2012 : 11:00 : JM Holyrood Code: OR54A562 Using Forward Monte-Carlo Simulation to Value American Barrier Options Dr Daniel Wei-Chung Miao (National Taiwan University of Science of Technology) and Dr Yung-Hsin Lee (National Taiwan University of Science and Technology) Monte-Carlo simulation is known to be an effective tool for European options pricing problems. However, its applications in the valuations of American options which have additional early exercise feature are more complicated and have relatively shorter history of study. The main issue of applying simulation to value American options is that the early exercise boundary (critical price) is unknown and therefore the simulation must use backward induction to estimate the critical price and determine whether the option should be exercised at each particular time. Among those methods based on backward induction, perhaps the most successful and prevailing one is the regression based method, usually termed the least square method (LSM), which was proposed by Longstaff and Schwartz (2001). The purpose of this study is to develop a forward Monte-Carlo method for the valuation of a number of types of American options. It is intended not to use backward induction as required by other methods. To achieve this, we introduce the ‘pseudo critical price’ which provides a wise way to determine whether a simulated stock price has entered the exercise region. We show that it provides exactly the same information as the ‘real critical price’ regarding early exercising an option but is much less computationally expensive. Since backward induction tends to take more time and give less accurate estimates, intuitively the proposed forward method may help achieve better efficiency and accuracy. In the first stage of this study, we have shown that the forward method works nicely for three types of the American style options, including vanilla (call and put), chooser and exchange options. In the second stage where the method is extended to value American barrier options, the special mathematical property near barrier poses more challenges. With a proper adaption we show that the ‘pseudo critical price’ can also be applied to the valuation of these exotic options, and the validity of the proposed method can be proved mathematically. Through a series of numerical examples where the forward method is tested against the standard LSM, we demonstrate that our method provides significant improvements in numerical efficiency and accuracy in contrast with the LSM. 04/09/2012 : 11:30 : JM Holyrood Code: OR54A699 Empirical Bayes Methods for Discrete Event Simulation Ms Shona Blair (University of Strathclyde), Prof Tim Bedford (University of Strathclyde) and Prof John Quigley (University of Strathclyde) A vast body of literature exists on the statistical analysis of Discrete Event Simulation (DES) output data, yet the potential use of Empirical Bayes (EB) methods has so far been neglected. EB procedures offer a structured and theoretically-sound framework for the pooling of data obtained across a set of populations to support inference concerning the parameters of an individual population. This often enables more efficient inference in situations which feature a repeated structure, providing that sufficient ‘similarity’ exists between component populations. It seems intuitively reasonable that such an approach may be beneficial in DES model analysis. In this talk, the results of a computational study investigating the application of EB procedures in the estimation of DES performance measures are presented. In particular, 117 the practical significance, as well as the benefits and limitations of the approach are discussed. Issues in implementation and likely directions for future exploration are also addressed. 04/09/2012 : 12:00 : Holyrood Code: OR54A732 Simulation in the Supply Chain Domain: Evaluating Modelling Approaches Mr Chris Owen (Aston Business School), Dr Pavel Albores (Aston Business School) and Dr Doug Love (Aston Business School) Supply chains are often complex dynamic networks which involve the integration, coordination and synchronisation of activities between different business entities and the transmission of both material and information. Simulation is a powerful technique for analysing and improving supply chain management (SCM) since it is suited to the particular challenges presented. Three main methods of simulation have been used in SCM modelling: System Dynamics (SD), Discrete Event Simulation (DES) and Agent Based Modelling (ABM). There is received wisdom on when and where to apply these techniques, but it lacks conceptual rigour and there is little guidance to help practitioners to understand the suitability of a given method to their problem. Likewise, there are very few ‘back-to-back’ studies of applying different simulation methods to the same problem to provide more rigour and clarity to the comparison. A set of supply chain problems ranging from strategic to operational are modelled ‘back-to-back’ using different modelling approaches and the models and modelling process compared. Case studies include: the modelling the decision making process for centralisation of procurement in a large construction organisation; the modelling of the ‘Bullwhip’ effect; and, the modelling and simulation of a global logistics chain. The case studies include the analysis of a number of key factors and how they might influence choice of approach including: the level of problem (strategic to operational); feedback representation; human decision making; and, model purpose. A number of interesting and surprising findings emerge concerning these approaches This practical ‘back to back’ modelling of different problems has led to findings which will be of interest to supply chain practitioners and academics alike since they clarify the areas where each technique is likely to be most valuable, and where it may have weaknesses and limits. 05/09/2012 : 11:00 : JM Holyrood Code: OR54A564 Explorative Research into Current Practice of Experimentation in Discrete Event Simulation Dr Kathryn Hoad (University of Warwick), Dr Tom Monks (Peninsula College of Medicine and Dentistry) and Dr Frances O'Brien (University of Warwick) Experimentation is arguably one the largest and most active research areas within discrete-event simulation. However, studies of discrete-event simulation practice report little transfer of this theory into real world application. This paper explores this gap. We conducted interviews with a small number of DES practitioners in order to inform the construction of a larger online survey exploring the current practice of experimentation in DES. This paper presents early results from this investigation. 05/09/2012 : 11:30 : JM Holyrood Code: OR54A582 Modelling Assembly Lines using Simulation Mrs Norhanom Awang (University Malaysia Pahang) and Prof Razman Mat Tahar (University Malaysia Pahang) Nowadays, firms are competing against time. Time based competition that involves reducing the time required to develop and produce products, puts additional pressures upon manufacturers to impose on-time delivery to meet customer at the right time, right place and the right quantity. This study focuses on modelling the assembly line associated with cycle time of the production process in selected automotive plant which consists of five sections named front under, body line, roof & floor, main line and fitting line. The processes will be measured to evaluate the cycle time performance and how it affected the productivity. The flowchart of the process will be modelled and simulated. The objective is to proposed minimal cycle time, minimise work-in-progress and reduce lead time. 118 05/09/2012 : 12:00 : JM Holyrood Code: OR54A592 Modelling and Simulation of Oil Palm Plantation Operation System Mrs Fazeeda Mohamad (University) and Prof Razman Mat Tahar (Universiti Malaysia Pahang) Modelling and simulation were known as a powerful problem solving tool especially when dealing with the behaviour of a complex system. One of the functional areas of using this method is in the oil palm industry by evaluating and improving the process performance as a way to improve the production throughput, capacity and productivity. The paper begins with the detailed process description of the upstream activities in the oil palm supply chain process especially during the transportation activities. The study focuses on modelling the oil palm plantation associate with lead time and utilisation of resources as the control of it is significant due to its influence on the quality of the fruits. The issue of machines breakdown during the operation affects the performance in this industry. Arena simulation software is used and the ‘what if’ analysis is applied. The operation alternatives performed and the summary of the model outcome will provide an interpretation and meaningful report to the management. 06/09/2012 : 08:30 : JM Holyrood Code: OR54A556 TUTORIAL: Conceptual Modelling for Simulation Prof Stewart Robinson (Loughborough University) Conceptual modelling is the abstraction of a simulation model from the real world system that is being modelled; in other words, choosing what to model, and what not to model. This is generally agreed to be the most difficult, least understood and most important task to be carried out in a simulation study. With reference to some example problems, we explore the problem of conceptual modelling. In particular, we define a set of terminology that helps us frame the conceptual modelling task, we discuss the role of conceptual modelling in the simulation project life-cycle, and we identify the requirements for a good conceptual model. A framework that may be helpful for carrying out and teaching effective conceptual modelling is presented. 06/09/2012 : 11:30 : JM Holyrood Code: OR54A630 Developing Parallelised Modelling Environments using GAMOV Dr Richard Boakes (University of Portsmouth) and Mr Gareth Toomey (DSTL) Within defence operational analysis there is a growing requirement for rapid analytical processes. To address part of this problem, development of the Generic Aggregator Model Valuator (GAMOV) toolset was commissioned as described by Glover and Toomey (2011). GAMOV reduces reliance on proven hard-coded simulations by enabling the rapid production of models built specifically for the purposes of a study. Novel and extant concepts can be quickly expressed in GAMOV through distinct data and functionality interfaces that enable models to be constructed through an approach akin to plug-and-play, following Sargent (2005). GAMOV’s components are deployed as interconnected web services; using a RESTful (Fielding, 2000) webbased architecture. GAMOV can therefore exploit advances throughout the web technology stack, including the nascent Semantic Web, enabling concepts within GAMOV to be described through sets of ontologies (Hofmann et al, 2011). This web-based approach to GAMOV also means that services can be parallelised to exploit benefits in High Performance Computing (HPC) and distributed computer architectures, complementing Poulter’s (2011) embarrassingly parallel replication and extending it to include algorithm competition. GAMOV is currently approaching full operating capability, whereupon the toolset may be exploited by the wider analytical community. Building upon the concepts described in previous work, this paper details how the flexibility of GAMOV’s architecture will enable analysts to engage with and exploit technological advances such as the Semantic Web and HPC; and highlights related research into detailed analysis of GAMOV output. 119 06/09/2012 : 13:30 : JM Holyrood Code: OR54A537 Validation of Simulation Models for Operational Control Mr Michael Leyer (Frankfurt School of Finance & Management) and Mrs Sophie Meilinger (KfW) Managing a company’s operations is a continuous cycle during which short-term actions are considered, implemented, reviewed and adjusted in order to achieve a high efficiency. Identifying the most promising option under a complex environment and ever-changing conditions, however, constitutes a challenge even to experienced managers. In such cases, business process simulation for operational control provides a method to quickly evaluate the impact of decisions under diverse future scenarios in an artificial computer environment. As a result, the best solution is identified whereas the lengthy trial-and-error process and its associated costs are reduced. All expected benefits, however, are fully absorbed if the simulation is poorly designed. Consequently, a measure of a model’s appropriateness ” formally known as validation ” is needed to differentiate helpful from erroneous models. However, models for operational control have to be set up in a short-term and are only used for a short and limited period. Thus, these models differ from traditional strategic simulation models and thus require a different approach for validation. The research presented proposes a generic method for validating simulation models for operational control. Based on the building process of short-term simulation, components with risk of being inappropriately defined or integrated are first identified and then matched with corresponding validation methods from literature. With respect to quantitative measures however, traditional forward-looking approaches require too much data and time for ephemeral simulation models that remain unaltered for a short time only. A new, backwards-looking approach is introduced and measured as model building validity. Finally, the most suitable and effective methods are merged into a sequence of validation activities best geared towards the needs of simulation models for operational control. 06/09/2012 : 14:00 : JM Holyrood Code: OR54A676 Agile Simulation: The Way Forward? Ms Frances Sneddon (SIMUL8 Corporation) As Chief Technology Officer of a simulation software company, I’ve seen many software development projects and simulation projects succeed - and some that just ticked the box. What defines success in both is delivering an answer that meets your customer’s real needs. Did your simulation answer the question your customer really needed answering or just the one they asked? If your software or simulation answers their real question, then it changes how they work and becomes a key part of their toolkit - this is your success measure. Software development teams all over the world have adopted ‚Agile‛ because they believe it’s the approach that gets them closest to this goal. It puts the customer at the centre of the development process - and I’ve witnessed how this radically changes a project. Are there lessons to be learned from Agile for delivering successful simulations? Could an agile simulation method deliver simulations of more value to customers? 06/09/2012 : 14:30 : JM Holyrood Code: OR54A713 Measuring the Effects of Supply Risks on Supply Chain Partners’ Inventory Policy: A Simulation Study Miss Mualla Gonca Yunusoglu (Dokuz Eylul University), Dr Hasan Selim (Dokuz Eylul University) and Dr A. Serdar Tasan (Dokuz Eylul University) A supply chain is a network of facilities and activities that procure raw materials from suppliers, use these materials to produce intermediate and finished products, finally deliver finished products to customers. In today’s global competition, suppliers, manufacturers and distribution centres gain competitive advantage by collaborating on a supply chain network. However, as a result of globalisation, supply chains and logistics activities increase in complexity. Consequently, high complexity results in several kinds of supply chain risks. In this study, the effect of supply risks on inventory policy is examined by using simulation. The aim of this study is to identify locations and levels of safety stocks that lead to low inventory costs for the entire supply chain. In addition, the effects of supply risk parameter levels on supply chain costs are analysed. The simulation environment designed in this study consists of four retailers having different demand patterns, a 120 warehouse and two suppliers, namely, a reliable supplier and an unreliable supplier. Specifically, the reliable supplier is completely reliable in terms of product quality and supply risks. However, the unreliable supplier is a capacitated supplier with product quality risk, disruption risk, and variable delivery lead time. The operation of supply chain is simulated for 500 days. In this regard, average total daily cost of entire supply chain is selected as a performance measure. Specifically, the performance values of the first 50 days of simulation runs are excluded from the analyses due to warm-up conditions. The results reveal that holding 30% safety stock at the warehouse provides a higher cost saving than the other alternatives. 121 Stochastic Inventory Control Organiser: Roberto Rossi 06/09/2012 : 08:30 : ST.L.Brewster Code: OR54A544 On Using Stochastic Dynamic Programming for Solving a Perishable Inventory Model Dr Eligius M.T. Hendrix (Wageningen University), Dr Rene Haijema (Wageningen University), Mrs Karin G.J. Pauls-Worm (Wageningen University) and Dr Roberto Rossi (University of Edinburgh) We consider a production planning problem over a finite horizon of T periods of a perishable product with a fixed shelf life of J periods. The demand is uncertain and non-stationary such that one produces to stock. To keep waste due to outdating low, one issues the oldest product first, i.e. FIFO issuance. A service level applies to guarantee that the probability of not being out-of-stock is higher than a fixed service level for every period. Any unmet demand is backlogged. The question dealt with here is whether stochastic dynamic programming can be applied to generate solutions up to a guaranteed accuracy. Therefore, we analyse the boundaries of the system and the behaviour for a deterministic model. In fact, solving the Bellman equations implies solving a global optimisation problem for each state value in a predefined grid. With increasing variance, the ranges of possible inventory levels increase and the function to be minimised gets smoother. In fact, as expected, the optimal order quantities decrease. We showcase the characteristics of the approach. 06/09/2012 : 09:00 : ST.L.Brewster Code: OR54A623 Forecasting Intermittent Demand by Hyperbolic-Exponential Smoothing Dr Steven Prestwich (University College Cork) Croston's method is superior to exponential smoothing when demand is intermittent, but in practice it has some drawbacks. Firstly, it assumes statistical properties of the data that may not be true. Secondly, it requires analysis to decide when it should be applied. Thirdly, it continues to forecast demand indefinitely when no further demands occur. We describe a new forecasting method called Hyperbolic-Exponential Smoothing that slowly decays when demand is zero, converges to single exponential smoothing on nonintermittent demand, forecasts as accurately as Croston's method on standard intermittent demand, and is almost unbiased on demand with stochastic, regular and auto-correlated intervals. 06/09/2012 : 11:30 : ST.L. St.Trinneans Code: OR54A554 Dynamic Programming and Simulation to Discover a New Order Policy for Fresh Products Dr Rene Haijema (Wageningen University) For a product with a fixed shelf life, we compute a cost optimal stock-age dependent ordering policy by Stochastic Dynamic Programming, assuming fixed ordering costs and unit outdating and shortage costs. By simulation of the optimal policy we derive stock level dependent rules, e.g. an (sd, Sd) rule with day (d) dependent parameter values. It appears that such rules may not cover the structure of an optimal policy well. A broader class of order policies is discovered by studying the simulation results. The new class policies is still stock age dependent and are thus easy to implement in practice. The SDP-Simulation approach provides thus improved order policies as well as a procedure for determining near optimal parameter values. 122 06/09/2012 : 12:00 : ST.L. St.Trinneans Code: OR54A626 The Value of Substitution in a Product Recovery System with Separate Markets Dr Sarah Marshall (University of Strathclyde) and Prof Tom Archibald (University of Edinburgh) Increasing legislative and societal pressures are forcing manufacturers to become environmentally conscious and take responsibility for the fate of their goods after they have been used by consumers. As a result, some manufacturers operate hybrid systems which produce new goods and recover used goods. Product recovery describes the process by which used products are returned to their manufacturers or sent to a specialised facility for recovery, before being sold on the original or a secondary market. In cases where the functionality of the new and recovered goods remain the same or similar, some consumers may be willing to substitute one good for the other, if their preferred good is out of stock. This paper will present a product recovery system in which newly produced goods and recovered goods are sold on separate markets, but can act substitutes for each other. A semi-Markov Decision Process formulation of this problem is presented and is used to obtain an optimal policy, which specifies production, recovery and substitution decisions. The model is used explore the properties of such a system, and in particular, the managerial implications associated with upward and downward substitution strategies are investigated. It is found that offering substitution can allow firms to increase their profit and in some cases also increase the proportion of demand satisfied (as measured by the fill rate). Having the option to offer substitution also impacts on the optimal policy, with production being performed less frequently when downward substitution can be offered, and recovery being performed less frequently when upward substitution can be offered. 123 Stochastic Scheduling & Dynamic Allocation Problem Organisers: Diego Ruiz-Hernandez 04/09/2012 : 11:00 : JM Duddingston Code: OR54A619 The Evaluation of Risk in Pharmaceutical Research: A Study of Current Models and Techniques and a Formulation of Two New Measures of Risk Miss Anne-Marie Oreskovich (Oxford University) and Prof John Gittins (Oxford University) Our aims with this research are to describe current methods of risk evaluation in pharmaceutical research, compare them with two proposed new methods, and to implement suitable methods in our new software which has been created to evaluate risk in pharmaceutical research. We first provide a review of recent articles written on risk evaluation in drug research. We then propose two new risk measures, state and prove relevant lemmas and conjectures regarding these measures, and then show how they can be utilised, demonstrating with examples. We also introduce our software, Optimising Pharmaceutical Research Resource Allocation (OPRRA), which is based on a stochastic scheduling and allocation model for pharmaceutical research. This includes a dynamic allocation or index policy for maximising profitability. Our work on risk, though more widely applicable, is being developed, and will be illustrated, in this context at the moment. Although we are not formally considering a bandit problem, our work uses similar ideas, and hopefully will be of interest to people working with bandit problems. In conclusion, we found that our two new risk measures provide information over and above what is given by current risk measures. More specifically, in our opinion, none of the current risk measures service pharmaceutical research planning objectives sufficiently. Our two new risk measures ameliorate this problem by providing additional insights into the financial risk of a drug project in a transparent, accurate, consistent, and specific way. 04/09/2012 : 11:30 : JM Duddingston Code: OR54A633 Inventory Optimization and Process Flexibility via Approximate Dynamic Programming Dr Chris Kirkbride (Lancaster University) Consider an inventory system in which inventories for M products are replenished by N production factories. Process flexibility in this context is that each factory is able to produce a subset of the M products which are held in inventory. Such production flexibility provides greater operational scope to manage fluctuations in inventory positions that result from stochastic demands on the system over a dedicated (one factory one product) system. A fully flexible system (each factory produces M products) offers the greatest amount of flexibility but will have comparatively large set-up and operational costs. Jordan and Graves (1995) and the literature that has followed have shown that limited flexibility (less than M products at each factory) under specific production plans (chaining) can gain improvements in the production capacity over a dedicated system that are close to that achieved by a fully flexible system. To take advantage of process flexibility and its associated production capacity improvements, the period to period decisions that specify production levels of the different products at each factory are key. Theoretically, a production policy that minimises inventory costs can be achieved via dynamic programming. However, for problems of realistic size this is not possible due to the so-called curse of dimensionality caused by both the high dimensional state space and production combinations. To circumvent such challenges we propose to deploy the techniques of approximate dynamic programming (ADP). We present the results of a preliminary investigation that analyses the performance of a selection of standard ADP algorithms in inventory systems of tractable size 124 for which the computation of an optimal policy is feasible. How these techniques will be extended to tackle realistic problems will be described. 04/09/2012 : 12:00 : JM Duddingston Code: OR54A669 Dynamic Pricing in Yield and Revenue Management Mr Michael Pearson (Edinburgh Napier University) The yield management problem we address is traditionally posed by the budget airlines. These airlines do not have protected classes but instead use dynamic pricing to determine allocation policy. This usually means that passengers who book early get discounted rates, while those booking closer to the date of the flight pay higher rates. We expand on earlier work with the newsvendor problem to employ a network equilibrium solution which takes account of both customer and airline objectives to formulate a dynamic solution to the pricing of airline tickets. We make use of phase plane analysis when identifying performance during two stages of the pricing strategy consistent with historical demand and pricing information. The first phase plane (endogenous) identifies an area of optimal profit and its association with customer satisfaction and airline performance measures by mapping error variability against the difference between seat allocation and customer demand. The second phase plane (exogenous) identifies optimal performance with regard to the global market place by mapping error variability against the sum of seat allocation and customer demand. The optimal solution strategy follows closely the solution of a stochastic differential equation set with price and booking closure as control variables. We illustrate the method with data from the airline industry. 04/09/2012 : 16:30 : ST.L.Brewster Code: OR54A624 Some Index Policies for Stochastic Machine Maintenance Problems with Imperfect Maintenance and Performance Shocks Dr Diego Ruiz-Hernandez (University College for Financial Studies) and Dr David Delgado-Gómez (Universidad Carlos III de Madrid) In Glazebrook, K., Ruiz-Hernandez, D. and Kirkbride, C. (2006) we established the indexability of a class of restless bandits designed to model machine maintenance problems in which maintenance interventions have to be scheduled to mitigate escalating costs as machines deteriorate, and to reduce the chances of a machine breakdown. In this paper we further develop the findings in our earlier work by presenting two families of examples for which explicit formulae for the Whittle index can be derived. In the first case, we relax the previous assumption that maintenance interventions are perfect and allow for some randomness in the active transitions. For the second case we relax the ‘right-skip free’ assumption of our first model and introduce ‘performance shocks’ that increase the deterioration state of the machine in big jumps. Numerical investigation evaluates the performance of Whittle’s index heuristic. 125 Speaking Authors Index First Name Fran Fran Norhafiza Kerem Kerem James Khaled Ahmed Mouhamad Shaker Babakalli Efthalia Liz Norhanom Federico Ricardo Sayara Tolga Gary Gary Anthony Valerie Julian Tessa Patrick Ian Daniel Graeme Shona Surname Ackermann Ackermann Ahmad Akartunali Akartunali Alexander Alhamad AlHinai Ali Agha Alkali Anagnostou Archibald Awang Barnabè Barros Beg Bektas Bell Bell Bellotti Belton Benavides Franco Berg Beullens Birchmore Black Blair Blair Stream Code OR CON PSM FORE OPTIM OPTIM PSM SCHED OPTIM LOGIST OPTIM OR&S OR CON SIMUL OR&S COMM INFOSYST GREEN PROJ PROJ CREDRISK GREEN PSM PSM LOGIST OR&S GREEN OR&S SIMUL Abstract Ref. OR54A601 OR54A579 OR54A640 OR54A664 OR54A685 OR54A579 OR54A591 OR54A729 OR54A607 OR54A729 OR54A659 OR54A731 OR54A582 OR54A595 OR54A605 OR54A568 OR54A539 OR54A668 OR54A707 OR54A734 OR54A738 OR54A725 OR54A541 OR54A691 OR54A586 OR54A663 OR54A644 OR54A699 126 Abstract Short Title Managing projects in an uncertain world Teaching Problem Structuring Methods Mining pattern behavior Radiation Treatment Planning Optimization for VMAT Stochastic Optimisation of Reliability Development Teaching Problem Structuring Methods A Genetic Algorithm Reilability and maintenance of Complex plant Reliability and Resilience in the Supply Chain Reilability and maintenance of Complex plant Data Analysis in the Water Industry Prioritising Asset Replacement Modeling Assembly Line using Simulation Strategic Management, Role Playing&System Dynamics CoP Approach to Evaluating Expert Support Preparing your data for Advanced Analytics The Bi-Objective Pollution-Routing Problem intellectual framework Holon Rich Pictures Models of financial distress in Chinese firms Climate Change   a challenge for MCDA? A decision making framework Rich Pictures A Generalised Multi-Echelon Inventory Theory Assessing Strategic Risk in the Public Sector Grocery Deliveries from Depot to Store Modelling Degradation Control & Monitoring Assets Empirical Bayes Methods for DES Page 97 103 49 88 86 103 115 89 73 89 92 98 118 94 41 69 57 105 106 44 58 103 102 73 95 54 92 117 First Name Richard John Sally Andrew Dave David Samantha Raffaella Fred Fred Chris John Sophie Sophie Sophie Marco J. Arturo Cagla Yanto Yanto Michael Wenlin Sabrina Pei-Yi Richard Emily Jonathan Naomi Neville Neville Gerard Chris Surname Boakes Bowers Brailsford Brint Buxton Buxton Buxton Calabrese Cameron Cameron Campbell Carney Carr Carr Carr Carvalho Castillo-Salazar Cergibozan Chandra Chandra Charlton Chen Cheng Conniss Cookson Crook Crouch Curtis Curtis de Zeeuw Dent Stream Code SIMUL HEA HEA OPERA MAI-Wkshps MAI-Wkshps FORE CREDRISK COMM PSM ANALY HORIZON MAI MAI-TecTaster MAI-TecTaster SCHED ROUT META OPERA OPERA COMM OPERA HEA SCHED OPTIM CREDRISK OR in Ed PSM PSM COMM PLE Abstract Ref. OR54A630 OR54A646 OR54A726 OR54A657 OR54A768 OR54A775 OR54A550 OR54A558 OR54A739 OR54A621 OR54A751 OR54A696 OR54A719 OR54A765 OR54A772 OR54A567 OR54A735 OR54A710 OR54A670 OR54A681 OR54A649 OR54A674 OR54A656 OR54A660 OR54A571 OR54A557 OR54A721 OR54A736 OR54A752 OR54A744 OR54A762 127 Abstract Short Title Developing parallelised modelling environments Balancing beds and theatres Modelling the dental workforce in Sri Lanka Modelling alternate bargaining Build your own pub! Build your own pub! pm Modelling & Forecasting Pharmaceutical Life Cycles Default probability of SMEs Engineering Tests of TASER Devices Schools of Thought Analysis (SOTA) R for Analytics The Ten Commandments of Horizon Scanning Speed Networking 08:30 - 09:30 Technique Taster - Bayesian Methods Technique Taster - Bayesian Methods A Look Ahead Heuristic For MOSP Route generator for WSRP A Hybrid GA for the TRP Altruistic Entrepreneurship Schumpeterian entrepreneurship PSM & project management in community regeneration Infection control with strategic healthcare worker Information Design in Health Communications Scheduling RAF Air Traffic Control Personnel Models for Air Cargo Revenue Management Survival analysis OR Outreach at British Airways Issue Identification at the Front-End of a Study NATO SAS-087 Presented by Neville J Curtis KEYNOTE:Modelling to support communities Challenges in energy system modelling Page 119 59 65 83 79 79 50 46 42 102 39 67 75 77 78 115 113 82 85 84 42 83 60 114 87 44 100 104 102 41 36 First Name Guido Guido Guido Patrick Orville Colin Richard Tracey Gunes Dafydd Masoud Javier Jason Kimon Anna James Deborah Pablo Narges Rene UngKyu Stephen Eligius M.T. Giles Giles Giles Giles Kathryn Penny Paul Paul Surname Diepen Diepen Diepen Driscoll D'Silva Eden Eglese England Erdogan Evans Fakhimi Faulin Field Fountoulakis Franceschetti Freeman Gee Gonzalez-Brevis Haghi Haijema Han Harwood Hendrix Hindle Hindle Hindle Hindle Hoad Holborn Holmstrom Holmstrom Stream Code MAI-TecTaster MAI-TecTaster OPTIM LOGIST HEA PROJ GREEN HEA LOGIST FORE HEA GREEN PLE OPTIM GREEN PROJ OR&S OPTIM ROUT STOCHINV OR&S OR&S STOCHINV MAI-TecTaster MAI-TecTaster MAI-TecTaster MAI-TecTaster SIMUL ROUT HEA HEA Abstract Ref. OR54A766 OR54A771 OR54A588 OR54A560 OR54A705 OR54A715 OR54A547 OR54A531 OR54A690 OR54A740 OR54A651 OR54A549 OR54A759 OR54A603 OR54A545 OR54A528 OR54A659 OR54A596 OR54A647 OR54A554 OR54A598 OR54A521 OR54A544 OR54A763 OR54A764 OR54A773 OR54A774 OR54A564 OR54A689 OR54A652 OR54A662 128 Abstract Short Title Technique Taster - Constraint Programming Technique Taster pm - Constraint Programming Solving Constraint Programming problems with AIMMS Acquisition & Acceptance Risks in Military Systems General practitioner funding formula KEYNOTE: Lessons for Project Management Fuel Emissions Optimization in VRPs Operational Research at the Welsh Government The Orienteering Problem with Variable Profits Anomaly Detection in Point Clouds Simulation for sustainability in healthcare Environmental Assessment of Transport Routes Operational Research - Supporting Defence and Secu Matrix-free Interior Point Method The Time-Dependent Pollution Routing Problem Cultural Diversity and Workplace Dynamics Data Analysis in the Water Industry On a warm-started primal-dual column generation me Lower + Upper Bounds for joint problem re Routing Discovering a new order policy Systems Thinking applied to Triple Helix Theories Is there still life in Stafford Beer's VSM? SDP in perishable inventory control Technique Taster - Rich Pictures Technique Taster - Systems Modelling Technique Taster pm - Systems Modelling Technique Taster pm - Rich Pictures Current Practice of Experimentation in D.E.S Methods for solving the dynamic PDPTW Melanoma pathways flight simulator Designing dementia care with group modeling Page 77 78 86 71 61 107 55 59 71 49 63 56 35 89 54 105 92 87 112 122 94 93 122 77 77 78 78 118 113 61 64 First Name John I-Hsuan John Susan Susan Kwei-Long Kenneth Kyunghyun Nasir Navid Jennifer Mohsen Mohsin Thomas Jooyoung Jose Nicholas Philip Angel A. Konstantinos Ruth Ruth Edward Ahmed Inna Eun-Ju Chris Vincent Nikoletta Christina Martin Chia-Wei Surname Holt Hong Hopes Howick Howick Huang Huh Hussain Izady Jackson Jafari Songhori Jat Jeffries Jeon Jimenez Jones Jones Juan Kaparis Kaufman Kaufman Kent Kheiri Kholidasari Kim Kirkbride Knight Koleri Konstantinidou Kunc Kuo Stream Code PSM HEA PLE OR CON PROJ HEA OR&S OR CON QUE COMM OPERA LOGIST HEA FORE ROUT OR CON OR&S GREEN OPTIM COMM MAI ROUT META LOGIST HEA STOCH OR in Ed FORE FORE OR&S OPERA Abstract Ref. OR54A518 OR54A730 OR54A760 OR54A601 OR54A715 OR54A688 OR54A639 OR54A682 OR54A698 OR54A693 OR54A741 OR54A611 OR54A593 OR54A570 OR54A572 OR54A687 OR54A665 OR54A532 OR54A589 OR54A616 OR54A718 OR54A638 OR54A697 OR54A634 OR54A666 OR54A633 OR54A575 OR54A683 OR54A540 OR54A524 OR54A677 129 Abstract Short Title OR for Air military assistance to Security and Dev Mechanism Design for Examination Resource Alloc. OR and analytics ” an opportunity for growth Managing projects in an uncertain world KEYNOTE: Lessons for Project Management Designing Quota Size and Patient Arrival Time Group dynamics and mental model convergence SWOT Analysis using General Morphological Analysis A Queueing Model for Specialty Clinics Researching ‘Wicked’ Issues - ‘Messy’ world of COR Complex Supply Network Time-differentiated service parts distribution Location & Scheduling of Blood Collection Sessions Wind power quantile forecasting Ant Colony Optimization for a school bus routing NHS provider landscaping Integrating Behaviour Change into OR The VRP with Multiple Driving Ranges The Stochastic Network Loading Problem OR in the 3rd sector MAI - Introduction and Welcome 08.20 - 08.30 Vehicle Routing with Dependent Vehicles Hyper-heuristics for Magic Squares Judgemental Adjustment in an Inventory System Sequential drug decision problems in healthcare Inventory optimization via ADP OR in Schools Long-term Energy Forecasting:The case of EU Forecasting Solvency of Greek Banks Generic models for start-ups Shelf-space competition Page 103 61 34 97 107 60 94 96 109 42 84 73 66 48 111 97 93 54 90 41 75 111 82 74 63 124 100 48 51 91 85 First Name Paul John Brian Michael Jiawei Zhiyong Paul Yan Emma Yan Emma Hannah Rasaratnam Altea MENG Mashael Jonathan Virginia Luisa Carol Sarah Nikolaos Fraser Ken Vadim Jose M. Daniel Wei-Chung Ian Fazeeda Jairo R John Fernando Michelle Arash Surname Lam Lamb Lehaney Leyer Li Li Liddiard Liu Liu Locke Logendran Lorenzo Arribas MA Maashi Malpass Marques Spiegler Marshall Marshall Mavroeidis McLeod McNaught Melnitchouk Merigo Miao Mitchell Mohamad Montoya-Torres Morecroft Moreira Morris Mostajeran Gourtani Stream Code ANALY META INFOSYST SIMUL SCHED CREDRISK ANALY OR&S OPERA HORIZON SCHED GREEN CREDRISK OPTIM OR CON LOGIST HEA STOCHINV FORE GREEN FORE CREDRISK FORE SIMUL OR CON SIMUL ROUT OR&S CREDRISK PROJ OPTIM Abstract Ref. OR54A743 OR54A548 OR54A712 OR54A537 OR54A533 OR54A636 OR54A667 OR54A675 OR54A672 OR54A745 OR54A559 OR54A614 OR54A627 OR54A714 OR54A701 OR54A602 OR54A695 OR54A626 OR54A584 OR54A542 OR54A750 OR54A546 OR54A700 OR54A562 OR54A684 OR54A592 OR54A572 OR54A535 OR54A692 OR54A650 OR54A578 130 Abstract Short Title Agile approach to knowledge discovery of web log Variable neighbourhood structures KM and sustainable quality Validation of simulation models for operational co A Local Search Heuristic to Compute Competitive TSP corporate governance in finance distress Royal Mail Goes To The Olympics! Competitive archetypes Impact of competition on capability development pa Conceptual Force Development Bicriteria Forge Scheduling Sustainability and seafood banquets, an utopia? UK SMEs through the 'Credit Crunch' GD based for HHMO Understanding Mobile Workforce Productivity Developing simplified and linearised models SoApt: Service Option Assessor and Prioritisation Product recovery with substitution Natural Gas Demand Forecasting Dynamic collection scheduling Predictive maintenance modelling with DBNs Stochastic parametric time to default model Decision making in complex environments Forward Monte-Carlo for American Options SSM for BRE Modeling and simulation Ant Colony Optimization for a school bus routing Metaphorical Models and Credible Worlds Joint credit losses and Poisson processes Planning with the OI Factor An MPEC approach to electricity trading strategy Page 40 81 70 120 115 45 39 91 84 67 114 56 46 86 98 72 63 123 51 55 49 45 51 117 99 119 111 94 44 107 90 First Name Penelope Urszula Ali Konstantinos Bjorn Frances Stephen anne-marie Louise Chris Rosane Christina Gu Alberto Michael Fotios Michael Steven Yi Robert Prasanna Kumar John John Vivienne Brian Jana Stewart Mark Patrick Geoff Diego Surname Mullen Neuman Niknejad Nikolopoulos Nygreen O'Brien O'Donnell oreskovich Orpin Owen Pagano Pagel Pang Paucar-Caceres Pearson Petropoulos Pidd Prestwich Qu Raeside Ramaswamy Ranyard Ranyard Raven Reddy Ries Robinson Roper Rose Royston Ruiz-Hernandez Stream Code HEA ROUT LOGIST FORE OPTIM OR&S PRES STOCH OR in Ed SIMUL PROJ HEA ROUT OR in Ed STOCH FORE HEA STOCHINV GREEN FORE ROUT MAI MAI GORS HEA META SIMUL MAI PRES PLE STOCH Abstract Ref. OR54A673 OR54A635 OR54A608 OR54A573 OR54A577 OR54A658 OR54A753 OR54A619 OR54A575 OR54A732 OR54A618 OR54A610 OR54A628 OR54A629 OR54A669 OR54A585 OR54A606 OR54A623 OR54A615 OR54A565 OR54A529 OR54A724 OR54A718 OR54A746 OR54A566 OR54A680 OR54A556 OR54A718 OR54A754 OR54A761 OR54A624 131 Abstract Short Title Over-analysis in Health Care and Health Services Receding Horizon Approach to Gate Allocation Two Phase Optimization in Reverse Logistics Forecasting of Policy Implementation Strategies Deployment of ICT Services in Cloud Data Centres Social media and engagement with scenario projects Modelling demand and performance in HMRC call cent The Evaluation of Risk in Pharmaceutical Research OR in Schools Simulation in the Supply Chain Domain Project Success: an ethical perspective Monitoring outcomes in paediatric heart surgery Variants of the Capacitated Arc Routing Problem Statistical Model to estimate Quality Research Dynamic Pricing in Yield and Revenue Management KEYNOTE: Forecasting most successful products KEYNOTE:OR and the industrialisation of healthcare Forecasting Intermittent Demand Green Intermodal Freight Transportation Unlocking the Potential of Computer Games Metaheuristics for Vehicle Routing Problems Pract.- Acad.Collab 16:30 + Review of day 17:45 MAI - Introduction and Welcome 08.20 - 08.30 GORS Recruitment MCDA approaches to prioritisation in NICE Instance-specific parameter tuning in practice TUTORIAL: On Conceptual Modelling for Simulation MAI - Introduction and Welcome 08.20 - 08.30 From Operational Research to Operational Planning Operational Research for the real world: big quest Index Policies for Stochastic Machine Maintenance Page 64 112 74 48 88 93 38 124 100 118 106 62 111 100 125 47 62 122 57 50 110 75 75 53 66 81 119 75 37 33 125 First Name Caroline Wendy L Tim Edward Terry Rob Jo Frances Frances Frances John Matthew Martin Kathryn Kemal Jerry Kate Kate Mohamed Ian Jacqui Jacqui Jacqui Eirini-Elisavet Nikolaos Christopher Kaoru Gareth Juan Pablo Miki Simon Ryunosuke Surname Sanders Schultz Schulze Seagriff Shone Smedley Sneddon Sneddon Sneddon Soje Soulby Spollen Stewart Subulan Swan Swatridge Swatridge Tadjer Taylor Taylor Taylor Taylor Theodorou Theodorou Tofallis Tone Toomey Torres Tsutsui Turner Usami Stream Code INFOSYST HORIZON SCHED PSM QUE INFOSYST MAI-Wkshps MAI-Wkshps SIMUL OR CON ROUT OR in Ed ROUT GREEN META MAI-Wkshps MAI-Wkshps HEA ANALY ANALY MAI-Wkshps MAI-Wkshps FORE FORE FORE OPTIM SIMUL OR&S OPTIM PRES FORE Abstract Ref. OR54A553 OR54A748 OR54A661 OR54A620 OR54A594 OR54A561 OR54A767 OR54A777 OR54A676 OR54A678 OR54A632 OR54A703 OR54A599 OR54A709 OR54A625 OR54A770 OR54A778 OR54A702 OR54A756 OR54A757 OR54A769 OR54A776 OR54A583 OR54A590 OR54A587 OR54A597 OR54A630 OR54A733 OR54A641 OR54A755 OR54A612 132 Abstract Short Title Event monitoring a large financial enterprise Systems, Scanning, and Scenarios Stochastic Unit Commitment by Column Generation Strategic Choice to Analyse Naval Deployment Price of Anarchy in Queueing Systems Weaving the innovation web Networking for Introverts Networking for Introverts pm Agile simulation ” the way forward? Decision analysis model of disease management Vehicle Routing Problem with driver familiarity Modelling latent preferences Dynamic congestion charging Green Tire CLSC Network Design via EI-99 Heuristic function resynthesis Personal and professional development Personal and professional development pm Service redesign Enter the Matrix Big Brother is Still Watching, but ... Data Visualisation Data Visualisation pm Forecasting short term electricity load demand Portfolio management with forecasting methods A comparison of relative accuracy measures Dynamic DEA with network structure Developing parallelised modelling environments Supporting internationalisation using SD GRS model in DEA Decision Analysis in Nuclear Decommissioning: Demand Forecast Considering Media Effects Page 69 68 116 103 109 69 78 79 120 96 112 101 110 57 81 78 79 65 39 39 79 79 52 52 49 89 119 91 88 38 47 First Name Martha César Dario Bastiaan Eliseo Miles Diederik David David David Bing Shu-Jung Sunny Jason Jiun-Yu Jiun-Yu Jiun-Yu Mualla Gonca Christos Surname Vahl Velandia Verhoef Vilalta-Perdomo Weaver Wijnmalen Worthington Worthington Wrigley Xu Yang Young Yu Yu Yu Yunusoglu Zikopoulos Stream Code COMM QUE CREDRISK COMM PROJ OR CON HEA QUE OR CON FORE LOGIST QUE HEA OPERA OPERA SIMUL LOGIST Abstract Ref. OR54A744 OR54A631 OR54A637 OR54A617 OR54A749 OR54A580 OR54A705 OR54A706 OR54A737 OR54A686 OR54A679 OR54A642 OR54A655 OR54A653 OR54A654 OR54A713 OR54A534 133 Abstract Short Title KEYNOTE:Modelling to support communities Co-ordinated queues at arterial intersections Aggregating Credit and Market Risk: Model Choice Mitigating potential conflicts in sustainability Methodology for Simulation Conceptual Modelling CB(RN) risk assessment multi-methodology General practitioner funding formula When to use Virtual Hold Technology Case studies in tender evaluation Evaluation of volatility forecasting models Supply chain competition and capacity constraints Hierarchical Queues Improving Performance of Emergency Medical Service Redesign and Innovate Aging Medical Institutes Integrated Healthcare Network Facilitators Effects of Supply Risks: A Simulation Study Remanufacturing with multiple collection sites Page 41 108 45 42 106 98 61 108 98 52 71 109 62 85 83 120 72 Delegate List First Name Fran Gareth Norhafiza Kerem Pavel James Khaled Mouhamad Shaker Babakalli Abdulaziz Efthalia Galina Stavros Thomas Liz Norhanom Federico Ricardo Rachel Charlene Sayara Tolga Gary Tony Valerie Tessa Neil Patrick Ian Daniel Gavin Graeme John Kathleen David Sally Andrew Samantha David Rosemary Raffaella Fred Chris John Sophie Marco Jose Arturo Cagla Sau Kwan Yanto Michael Surname Ackermann Adams Ahmad Akartunali Albores Alexander Alhamad Ali Agha Alkali Alnutaifi Anagnostou Andreeva Apostolou Archibald Archibald Awang Barnabè Barros Basnett Beckford Beg Bektas Bell Bellotti Belton Berg Berry Beullens Birchmore Black Blackett Blair Bowers Bowie Boyes Brailsford Brint Buxton Buxton Byde Calabrese Cameron Campbell Carney Carr Carvalho Castillo Salazar Cergibozan Chan Chandra Charlton Email Address fran.ackermann@strath.ac.uk izhafiza@yahoo.com KEREM.AKARTUNALI@STRATH.AC.UK p.albores@aston.ac.uk james.alexander.100@strath.ac.uk khabsigns@hotmail.com ali.mouhamad-shaker@strath.ac.uk babakalli.alkali@gcl.ac.uk abdulaziz.nutaifi@aramco.com touloubaki@hotmail.com Galina.Andreeva@ed.ac.uk stavros.apostolou@gmail.com tarchibald@ed.ac.uk liz.archibald@decisionlab.co.uk anumjp@yahoo.com barnabe@unisi.it r-barros@uniandes.edu.co rachel.basnett@sellafieldsites.com charlene.beckford@sellafieldsites.com sayara@datanut.co.uk T.Bektas@soton.ac.uk bellgaa@lsbu.ac.uk a.bellotti@imperial.ac.uk val.belton@strath.ac.uk tb79@hw.ac.uk P.Beullens@soton.ac.uk phd10ib@mail.wbs.ac.uk dan.black@ed.ac.uk gavin.blackett@theorsociety.com g.a.blair@hotmail.com j.a.bowers@stir.ac.uk david.boyes@dwp.gsi.gov.uk s.c.brailsford@soton.ac.uk A.Brint@sheffield.ac.uk samantha.buxton87@gmail.com david.buxton@decisionlab.co.uk raffaella.calabrese1@unimib.it Fred.Cameron@cogeco.ca ccampbell@mango-solutions.com sophie@baysconsulting.co.uk mamc@iceb.ufop.br psxjaca@nottingham.ac.uk cagla.cergibozan@deu.edu.tr Y.Chandra@leeds.ac.uk m.charlton@shu.ac.uk 134 First Name Phillip Wen-Chih Pei-Yi Neil Richard Christina Emily John Jonathan Naomi Sonya Neville Alec Roy Chris Guido Patrick Robert Colin Phil Richard Mark Tracey Gunes Dafydd Masoud Javier Jason Gregor Kimon Anna James Ramune Deborah Pablo Arash Gillian Narges Rene UngKyu Stephen Eligius M.T. Katy Penny Paul John I-Hsuan John Susan Kwei-Long Kenneth Kyunghyun Nasir Surname Chatikobo Chen Cheng Cochran Conniss Constantinidou Cookson Crocker Crook Crouch Crowe Curtis Davidson Davy Dent Diepen Driscoll Dyson Eden Egert Eglese Elder England Erdogan Evans Fakhimi Faulin Field Finlayson Fountoulakis Franceschetti Freeman Gedgaudaite Gee Gonzalez-Brevis Gourtani Groom Haghi Haijema Han Harwood Hendrix Hoad Holborn Holmstrom Holt Hong Hopes Howick Huang Huh Hussain Email Address phillip.chatikobo@scottishwater.co.uk wenchih@faculty.nctu.edu.tw pycheng555@gmail.com neil_cochran@hotmail.com psxrc@nottingham.ac.uk ckonstantinidou@fsu.gr e.cookson@lancaster.ac.uk john.crocker@o-sys.com j.n.crook@ed.ac.uk naomi.crouch@ba.com sonya.crowe@ucl.ac.uk neville.curtis@dsto.defence.gov.au Roy.Davy@ScottishWater.co.uk chris.dent@durham.ac.uk guido.diepen@aimms.com robert.dyson@wbs.ac.uk colin.eden@strath.ac.uk r.eglese@lancaster.ac.uk mark.e@simul8.com smatje@cf.ac.uk G.Erdogan@soton.ac.uk evansd8@cf.ac.uk masoud.fakhimi@gmail.com javier.faulin@unavarra.es JRFEILD@DSTL.GOV.UK K.Fountoulakis@sms.ed.ac.uk jim.freeman@mbs.ac.uk gedgaudaiter@gmail.com deborah.gee@scottishwater.co.uk P.Gonzalez-Brevis@sms.ed.ac.uk arash.gourtani@gmail.com G.Groom@soton.ac.uk nh2w07@soton.ac.uk Rene.Haijema@wur.nl ungkyuhan@gmail.com stephen.harwood@ed.ac.uk eligius.hendrix@wur.nl kathryn.hoad@wbs.ac.uk holbornpl@cf.ac.uk ph@holmstrom.se jholt@dstl.gov.uk ihong@ntu.edu.tw jhopes@uk.ey.com susan.howick@strath.ac.uk craighuang@ntu.edu.tw phd09kh@mail.wbs.ac.uk hussain@strategyforesight.org 135 First Name Navid Jennifer mohsen Mohsin Nasir Thomas Jooyoung Jose Philip Sean Nicholas Ángel Alejandro Konstantinos Ruth Michelle Edward Martin Ahmed Inna Eunju Christopher Vincent Nikoletta Niraj Martin Wasakorn Paul John Brian Michael Jiawei Zhiyong Paul Alex Akrivi Kuangyi Hannah Rasaratnam Altea Meng Mashael Jon Virginia Luisa Carol Sarah Nikolaos Richard Sean Andrew Fraser Ken John Vadim Surname Izady Jackson jafari songhori Jat Jefferies Jeon Jimenez Jones Jones Jones Juan Kaparis Kaufman Kennedy Kent Keys Kheiri Kholidasari Kim Kirkbride Knight Koleri Kumar Kunc Laesanklang Lam Lamb Lehaney Leyer Li Li Liddiard Lisikovs Litsa Liu Locke Logendran Lorenzo Arribas Ma Maashi Malpass Marques Spiegler Marshall Marshall Mavroeidis Maxwell McCann McGee McLeod McNaught Medhurst Melnitchouk Email Address n.izady@soton.ac.uk jjackson@lincoln.ac.uk mj2417@gmail.com lixmnja@nottingham.ac.uk tj64@kent.ac.uk jooyoung.jeon@strath.ac.uk jose_fernando_jimenez@hotmail.com prjones@dstl.gov.uk sean.jones@nats.co.uk nick.m.jones@uk.pwc.com suport_admreecrca@uoc.edu k.kaparis@lancaster.ac.uk ruth.kaufman@btinternet.com psxerk@nottingham.ac.uk keysrus@btinternet.com axk@cs.nott.ac.uk i.kholidasari@edu.salford.ac.uk ejkim1023@gmail.com c.kirkbride@lancaster.ac.uk knightva@cf.ac.uk koleri.nikoletta@gmail.com n.kumar@sheffield.ac.uk martin.kunc@wbs.ac.uk psxwl3@nottingham.ac.uk paul.lam@forward.co.uk j.d.lamb@abdn.ac.uk b.lehaney@qaa.ac.uk m.leyer@fs.de jwl@cs.nott.ac.uk zhiyong.li@ed.ac.uk paul.a.liddiard@royalmail.com alex@credit-scoring.co.uk akrivilitsa@yahoo.gr kuangyi.liu@uk.pwc.com hllocke@dstl.gov.uk logendrr@engr.orst.edu altealorenzo@gmail.com M.Ma-4@sms.ed.ac.uk mvm@cs.nott.ac.uk jonathan.malpass@bt.com MarquesVL@cardiff.ac.uk cm87@stir.ac.uk sarah.e.marshall@strath.ac.uk nikos2mb@hotmail.com richard.maxwell@btinternet.com sean.mccann@sellafieldsites.com andrewmcgee@netscape.net fnm@soton.ac.uk k.r.mcnaught@cranfield.ac.uk john@larrainzar.co.uk vmelnich@gmail.com 136 First Name Jose M. Daniel Wei-Chung Ian Fazeeda John Fernando Michelle Penelope Robert Urszula Ali Bjorn Frances Tony Stephen Anne-Marie Louise Djamila Jamal Christopher Daniel Rosane Christina Gu Alberto Michael Fotios Dobrila Sanja Michael Steven Rosalind Yi Robert John Vivenne Brian Catrin Stewart Roberto Geoff Diego Graham John Desmond Caroline Wendy L Tim Alistair Edward Alison Rob Anusua Surname Merigo Lindahl Miao Mitchell Mohamad Morecroft Moreira Morris Mullen Murray Neuman Niknejad Nygreen O'Brien O'Connor O'Donnell Oreskovich Orpin Ouelhadj Ouenniche Owen Owen Pagano Pagel Pang Paucar Pearson Petropoulos Petrovic Petrovic Pidd Prestwich Pyne Qu Raeside Ranyard Raven Reddy Roberts Robinson Rossi Royston Ruiz-Hernandez Russel Ryan Sanders Schultz Schulze Scotland Seagriff Shimmin Shone Singh Roy Email Address jose.merigolindahl@mbs.ac.uk miao@mail.ntust.edu.tw ian.mitchell@bis.gsi.gov.uk adeezaf@yahoo.com jmorecroft@london.edu fernandoufmg@hotmail.com m.morris@mmu.ac.uk penelope.mullen@btinternet.com robert.murray@scottishwater.co.uk uxn@cs.nott.ac.uk niknejas@coventry.ac.uk bjorn.nygreen@iot.ntnu.no Frances.O-Brien@wbs.ac.uk tony.o'connor@dh.gsi.gov.uk steve.o'donnell@hmrc.gsi.gov.uk anne-marie.oreskovich@magd.ox.ac.uk louise.orpin@theorsociety.com djamila.ouelhadj@port.ac.uk Jamal.Ouenniche@ed.ac.uk c.owen@aston.ac.uk dan.james.owen@googlemail.com r.pagano@mmu.ac.uk c.pagel@ucl.ac.uk gu.pang@ncl.ac.uk a.paucar@mmu.ac.uk m.pearson@napier.ac.uk fotpetr@gmail.com csx200@coventry.ac.uk sanja.petrovic@nottingham.ac.uk m.pidd@lancaster.ac.uk s.prestwich@cs.ucc.ie r.pyne@palgrave.com vanilla_0103@hotmail.com r.raeside@napier.ac.uk jranyard@cix.co.uk vivienne.raven@hmrc.gsi.gov.uk brian.reddy@gmail.com catrin.roberts@dft.gsi.gov.uk s.l.robinson@lboro.ac.uk roberto.rossi@ed.ac.uk geoff.royston@gmail.com d.ruiz@cunef.edu ryanjohn@btinternet.com csanders@westpac.com.au wendy@infinitefutures.com t.schulze-2@sms.ed.ac.uk alistair.scotland@strath.ac.uk eseagriff@dstl.gov.uk alison.c.shimmin@sellafieldsites.com rshone@hotmail.com A.SinghRoy@napier.ac.uk 137 First Name John Jo David Nei Yoshihiro Matthew Martin Kathryn Andrew Kemal Jerry Kate Mohamed Ian Jacqui Nikolaos Eirini-Elisavet Richard Christopher Kaoru Gareth Juan Miki Amanda Simon Ryunosuke Martha Cesar Bastiaan Eliseo Jonathan Miles Daniel Diederik Mik David David Bing Shu-Jung Sunny Jason Jiun-Yu Mualla Gonca Christos Surname Small Smedley Smith Soma Soulby Spollen Stewart Stothers Subulan Swan Swatridge Tadjer Taylor Taylor Theodorou Theodorou Thompson Tofallis Tone Toomey Torres Tsutsui Tucker Turner Usami Vahl Velandia Verhoef Vilalta-Perdomo Warwick Weaver Welsh Wijnmalen Wisniewski Worthington Wrigley Xu Yang Young Yu Yunusoglu Zikopoulos Email Address jds340@gmail.com jo.smedley@newport.ac.uk davidandtina@endfield.org.uk soma@ita.br psxms6@nottingham.ac.uk Martin.Spollen@sibni.org k.stewart@napier.ac.uk kemal.subulan@deu.edu.tr jsw@cs.stir.ac.uk kate.swatridge@decisionlab.co.uk m.tadjer@my.westminster.ac.uk ian.taylor@flyingbinary.com jacqui.taylor@flyingbinary.com nikos.t.theodorou@gmail.com elizabeth3487@gmail.com richard.a.thompson@sellafieldsites.com c.tofallis@herts.ac.uk tone@grips.ac.jp GDTOOMEY@mail.dstl.gov.uk j.p.torres@warwick.ac.uk ATucker@dstl.gov.uk simon.d.turner@magnoxsites.com dm106203@cc.seikei.ac.jp martha@cict.demon.co.uk psxcdv@nottingham.ac.uk bastiaan.verhoef@rbs.com evilaltaperdomo@lincoln.ac.uk warwick@lsbu.ac.uk m.weaver@napier.co.uk daniel.welsh@ba.com diederik.wijnmalen@tno.nl mik@mikwisniewski.com d.worthington@lancaster.ac.uk david.wrigley@ntlworld.com b.xu@rgu.ac.uk sunnyy@essex.ac.uk Jas_young@live.com jyyu@ntu.edu.tw gonca.yunusoglu@deu.edu.tr cziko@auth.gr 138 WALKING ROUTES IN HOLYROOD PARK YOU ARE HERE! 139 JOHN McINTYRE CONFERENCE CENTRE GROUND FLOOR ENTRANCE FIRST FLOOR PLAN Salisbury Stairs and lift up from entrance lobby Holyrood Registration Exhibition Concourse Centro Bar Pentland West Restaurant Prestonfield Pentland East Duddingston ST LEONARD’S HALL Nelson Room St Trinnean Room (Drinks reception on Wednesday 5 Sept) STAIRS TO FIRST FLOOR Pollock Room Cowan Room Brewster Room MAIN ENTRANCE Bonnar Room for meals (not Gala Dinner) on GROUND FLOOR CONFERENCE CAMPUS MAP OR SOCIETY REGISTRATION ACCOMMODATION EDINBURGH FIRST CHECK-IN AND CHECK-OUT Chancellor’s Court SESSIONS, TEA & COFFEE, MEALS (Not Gala Dinner) CASH POINT & SHOP Reception Centre SESSIONS St Leonard Hall John McIntyre Centre South Hall GALA DINNER