The control of today and prediction of the future using predictive production scheduling Your logo Wim De Bruyn lecturer ICT, R&D FBO, UCG product mgt. Inxites Bert Van Vreckem Lecturer ICT, FBO, UCG Researcher Prinsyslab, UCG SAPience.be User Day 2012 1 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 2 Introduction Your logo Predictive = What if we try ..... SAPience.be User Day 2012 3 ISA-95 Functional enterprise control model SAPience.be User Day 2012 Your logo 4 ISA-95 Production Information Overlap SAPience.be User Day 2012 Your logo 5 ISA-95 Production Segment Capabilities SAPience.be User Day 2012 Your logo 6 ISA-95 Product Definition Model SAPience.be User Day 2012 Your logo 7 ISA-95 Process segment relations Connecting a production request with a routing SAPience.be User Day 2012 Your logo 8 Current and future Production capabilities SAPience.be User Day 2012 Your logo 9 ISA-95 Production Schedule Model SAPience.be User Day 2012 Your logo 10 Short time Planning overview Your logo • GANTT chart, connecting resources with production orders (same colour) SAPience.be User Day 2012 11 Production Scheduling is … Your logo • telling a production facility when to make, with which staff, and on which equipment. • allocation of jobs to scarce resources • a combinatorial optimization problem • maximize and/or minimize objective(s) • subject to constraints SAPience.be User Day 2012 12 Production Scheduling Your logo • shorten delivery times • increase variety in end-products • shorten production lead times • increase resource utilization • improve quality, reduce WIP • prevent production disturbances (machine breakdowns) More products in less time! Less cost! More profit! Lower ecological impact! SAPience.be User Day 2012 13 Production Scheduling in Manufacturing Planning Framework Your logo • Long range prediction and sales planning • Facility and resources planning • Demand management, aggregate and workforce planning • Order acceptance and resource loading • Shop floor scheduling, workforce scheduling SAPience.be User Day 2012 14 Structure of APS SAPience.be User Day 2012 Your logo 15 Consumer goods planning Your logo SAPience.be User Day 2012 16 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 17 Ok! ERP-system Information Transfer Information Transfer Efficiency Criteria Numerical Data Production Process Description Information Transfer Information Transfer Optimization Objective(s) Variable Values Constraints Information Transfer Database ISA-88 or ISA-95 Compliant and Complete 𝐹 = 180 Good Schedule Production Schedule 𝐹 = 150 Automatic Parameter Adjustment Optimization Algorithm Compliant? Complete? Scheduler/ Automatic Decision Maker Manual Parameter Adjustment Constraints Consistency Check subject to 22 March 2012 Your logo Planner/ Decision Maker UCG - INXITES R&D Inconsistent? 𝑛 𝑎𝑖 𝑥𝑖 ≤ 𝑏 → ∅ 𝑖=0 18 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 19 Perfect Plant implementation (B2MML to SAP at Polar © 2004 World Batch Forum) SAPience.be User Day 2012 Your logo 20 Mapping SAP PP-PI, ISA95 Production Schedule, ISA 88 and the Physical Model Your logo B2MML to SAP at Polar © 2004 World Batch Forum SAPience.be User Day 2012 21 Simplified Schedule Request and Reponse Example MES SAP ME SAP MII B2MML Production Schedule XML (Request) SAP BC Your logo SAP PP PI B2MML Production Schedule XML (Response) The schedule can be refined and adapted in the MES execution part Netweaver interface Web service SAPience.be User Day 2012 22 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 23 Multiobjective Scheduling Your logo 𝑃𝑟𝑜𝑓𝑖𝑡 = 𝐶𝑥 → 𝑚𝑎𝑥 𝐶𝑜𝑠𝑡 = 𝐷𝑥 → 𝑚𝑖𝑛 𝐸𝑐𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝐼𝑚𝑝𝑎𝑐𝑡 = 𝐸𝑥 → 𝑚𝑖𝑛 SAPience.be User Day 2012 24 Multiobjective Scheduling Your logo subject to constraints 𝐴𝑥 ≤ 𝑏 Capacity Processing Times Man Power SAPience.be User Day 2012 Idle Times 25 Scheduling models Your logo • Many models, suitable for specific production processes • Continuous: Flow Shop Scheduling • Discrete: Open/Job Shop Scheduling • Batch: complex, several models depending on characteristics • Not for production scheduling: project scheduling, timetabling, ... SAPience.be User Day 2012 26 Scheduling Algorithms Your logo General Techniques: • Mathematical programming • linear, non-linear, (mixed) integer programming • Analytical (Exact) methods (enumeration) • branch-and-bound, branch-and-cut • dynamic programming • constraint satisfaction • Heuristics and meta-heuristics • genetic algorithm • tabu search • Artificial Intelligence • Reinforcement Learning • Hybrid Algorithms SAPience.be User Day 2012 27 Scheduling Algorithms (cont.) Your logo • Decomposition Techniques • Temporal decomposition (rolling horizon approach) • Machine decomposition (Shifting Bottleneck) • Dantzig-Wolf • MILP-decomposition SAPience.be User Day 2012 28 Scheduling Algorithms: Complexity Your logo • Analytical techniques • = algorithms that guarantee optimal solution • often infeasible • too many solutions (“NP-hard”) • mostly suitable for theoretical study of scheduling problems SAPience.be User Day 2012 29 Scheduling Algorithms: Complexity Your logo • “Non-analytical” techniques • = no guaranteed optimum, but feasible in time • paradigms • heuristics: use expert knowledge (“rules of thumb”) to create good schedules • meta-heuristics: simulated annealing, tabu search, genetic algorithms (cfr. local search) • artificial intelligence: rule-based, agentbased, expert systems • hybrid: combination of paradigms SAPience.be User Day 2012 30 Optimization Techniques: properties Your logo • Quality of Solutions Obtained (How Close to Optimal?) • Amount of CPU-Time Needed (Real-Time on a PC?) • Ease of Development and Implementation (How much time needed to code, test, adjust and modify) • Implementation costs (Expensive third-party components required?) SAPience.be User Day 2012 31 Software Solutions w.r.t. Optimization Techniques Your logo Implementation costs (Expensive LP-solvers required? Easy to implement?) Required solution quality? (Is an immediate answer required, or are long calculations allowed? Does customer accept complex solutions?) SAPience.be User Day 2012 32 How to reduce # searches? Your logo Dispatching Rules Value Objective Function Local Search Beam Search Branch and Bound CPU - Time SAPience.be User Day 2012 33 Decision Support Systems Your logo Important issues in design of DSS: • Database design and management • Data collection (e.g. barcoding system) • Module Design and Interfacing • GUI Design (Gantt-charts, etc.) • Design of link between GUI and algorithm library (data organization before transfer) • Internal Re-optimization • External Re-optimization SAPience.be User Day 2012 34 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 35 Real Case Chemical Batch Production Process Your logo (D.Borodin, B. Van Vreckem, W. De Bruyn, MISTA 2011 Scheduling Conference Proceedings) Big Seed Fermentation(2) Main Fermentation(5) Buffer tanks (4) Task Recovery (1) • Optimize the Production Process Objective • Minimize Total Tardiness (Customer Due-Dates) Solution Approach • Exact Optimal Solutions vs. Two Heuristic Methods SAPience.be User Day 2012 36 Real Case Results Comparison Problem Instance Exact Solution Your logo KL best KL time GA best GA time N10_1 90 91 7 93 5 N10_2 30 35 16 30 17 N10_3 42 56 10 44 6 N10_4 49 52 14 50 5 N10_5 43 48 6 45 12 N15_1 73 77 80 76 25 N15_2 43 45 112 45 34 N15_3 57 70 39 66 75 N20_1 52 54 490 54 180 N20_2 58 66 260 64 194 N30_1 _ 180 1304 186 1560 SAPience.be User Day 2012 37 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 38 GUI’S should allow: Your logo • Interactive Optimization • Freezing Jobs and Re-optimizing • Creating New Schedules by Combining Different Parts from Different Schedules • Cascading and Propagation Effects After a Change or Mutation by the User, the system: • does Feasibility Analysis • takes care of Cascading and Propagation Effects, • does Internal Re-optimization SAPience.be User Day 2012 39 GUI for Production Scheduling • • • • Your logo Gantt Chart Interface Dispatch List Interface Time Buckets (resource capacity loading) KPI dashboards SAPience.be User Day 2012 40 GUI: KPI dashboard Your logo Dashboard provides at-a-glance views of key performance indicators (KPIs) relevant to a particular objective, production or business process: capacities load, costs, profit, ecological impact, sales, marketing, human resources, etc. SAPience.be User Day 2012 41 GUI Important Objectives = KPIs? Your logo • Yes and No! • Due Dates (KPI or objective?) • Late orders • Maximum lateness • Average lateness, tardiness, earliness-tardiness, makespan • Productivity and Inventory Related (KPI or objective?) • Total Setup Time • Total Machine Idle Time • Resource usage (KPI or objective?) • Resource Shortage SAPience.be User Day 2012 42 Overall KPI concept Your logo • KPI-driven Production • Operations Research Approach: KPI-driven factory = KPIs as objective functions • Overall KPI: 𝑂𝑣𝑒𝑟𝑎𝑙𝑙𝐾𝑃𝐼 = 𝐶𝑥 + 𝐷𝑥 − 𝐸𝑥 + ⋯ − 𝑍𝑥 → min 𝑜𝑟 𝑡𝑎𝑟𝑔𝑒𝑡 𝑣𝑎𝑙𝑢𝑒 where 𝐶𝑥, 𝐷𝑥, … are various KPIs, + 𝑎𝑛𝑑 − mean maximization and minimization of a certain KPI • Goal: achieve production predictability, lean manufacturing SAPience.be User Day 2012 43 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 44 Integration from shopfloor to topfloor Your logo V.P. Mfg Mgr. Mgr. Mgr. SAP MII DCS / PLC MES Plant Historian SCADA / HMI LIMS/ Inspection / Equipment Testing Plant DB Environmental Building Management Plant Data Collection SAPience.be User Day 2012 Wireless Integration 45 SAP MII as KPI dashboard and and production schedule interface SAP MII extracts data from SAP ERP and provides real-time visibility and distribution to Plant Floor Systems of: SCM ERP Your logo SAP MII’s ability to perform transaction execution into SAP also enables automated, plant-level creation of: PLM Global Coordination PLAN MAKE DELIVER Manufacturing Integration & Intelligence Business Logic Services •Planned Orders •Bills of Material •Production & Process Orders •Material Inventory Levels •Inspection Lots Data •Master Recipes •Material Details •Batch Details •Resources & Functional Locations •Maintenance Work Order & Notification details •Material & Order Costs Visualization Composition Environment Quality Engine KPIs / Metrics / Alerts Data Services Simplified Execution Plant Historian DCS / PLC via OPC MES 22 March 2012 •Production Confirmations •Process Messages •Material Receipts •Material Consumptions •Material transfers •Inspection results recording •Quality Notifications •Batch Characteristic recording •Work Orders & results recording •Maintenance Notifications LIMS/ Inspection / Equipment Testing MAKE SAPience.be User Day 2012 Plant DB Environmental Building Management SCADA / HMI Wireless Integration 46 Better Asset usage, right product at the right time, less inventory, less waste “ Your logo Reduce throughput times 30% Reduce inventory 15 – 20% VP Mfg 14% less errors in production Quartile 1: 95% OEE vs. 78% Average* Reduce data capture efforts 65% Source: MESA International ! Plant Mgr. Machine Uptime 99.5% Qtr 1 vs. 93.6% Average* Qtr. 1: 98% First pass quality vs. 75% Avg.* * Benchmarks from ASUG Manufacturing Benchmarking Study Qtr. 1: 98.5% On-Time Delivery vs. 89.1% Avg.* SAPience.be User Day 2012 47 Interaction between SAP APO and Workcenter schedule Your logo Interactive Workcenter Schedule: Production Schedule updated every 30 minutes. Double Clicking on a production order provide the confirmation screen to enter new production. SAPience.be User Day 2012 48 More control – more interactivity – More planner and operator responsibility Alert ! Dashboards für die intelligente Fertigung SAP BI Weitere SAP-Lösungen SAP NETWEAVER Your logo Invoking of scheduling solution (SAP SCM APO). Machine disruption is considered as machine-breakdown in scheduling board. Finding an alternative capacity (manually or through rescheduling run). SAP Manufacturing (mySAP ERP) SAP MII Manufacturing Intelligence Manufacturing Integration Planner is able to respond to disruption in realtime and to resolve the conflict. SAPience.be User Day 2012 49 Agenda • • • • • • • • • • Your logo Introduction ISA-95: current and future production management Production Scheduling ISA-95 and Scheduling: Are they a lovely couple? ISA-95+Scheduling+SAP = ? Optimal Schedule and Algorithms The power of Algorithms on a Real Production Case From Algorithms to Management View: KPI dashboard Integration from Shopfloor to Topfloor Conclusions SAPience.be User Day 2012 50 Conclusions Your logo Optimization of Production Scheduling (Algorithms) implemented in the Standardized Environment (ISA-95compliant) and Incorporated in the Production Automation System (SAP) that allows visibility and transparency for all stakeholders involved in Production Process (KPI dashboard), will enable to: Reduce Cost Increase Profit Respect the Environment = Optimize Ecological Footprint SAPience.be User Day 2012 Realise Lean Production: avoid waste and timeloss 51 Thank you! Your logo SAPience.be User Day 2012 52