Operations Research Training for Regional Business Analysts The advent of the new TxDOT regions creates a myriad of business optimization opportunities. In-house operations research/management science (OR/MS) training for lead business analysts in the new regions will enhance their ability to deal with the wide variety of operational problems they will encounter. The ultimate result will be a cadre of trained lead analysts who can work synergistically, effectively direct the work of their staff analysts, and call upon external expertise as appropriate. Regional business analysts will be expected to assist managers with forecasting, resource allocation, cost minimization, efficiency analysis, and optimization of operations in general. A variety of tools are available to solve these classic business problems within the field of operations research/management science (OR/MS). The Institute for Operations Research and Management Science (INFORMS) defines their field as “the science of better.” For this reason, the department needs analysts with some degree of formal training in management science methods. While the universities (usually departments of Business, Industrial Engineering, or Mechanical Engineering) offer strong expertise in OR/MS, the department needs internal analysts with sufficient knowledge to select suitable researchers, provide guidance, and assess their work. The purpose of this training is to provide analysts with an enhanced capacity to recognize business problems that are amenable to established OR/MS methods. This is how the subject is typically presented in college management science courses. Once a situation is identified as a possible candidate for optimization methods (for example), a lead analyst can discuss the problem with other analysts and possibly consult with an established pool of university operations research experts. The training will be provided through a series of short courses held in Austin by selected university professors who are part of the OR/MS “brain trust” that has historically consulted to the department. Coursework will consist of a series of business decision case studies analyzed on a common Excel framework using standard features and some add-ons. Guided by instructors, each case will be worked through a “hands-on” approach by attendees in a TxDOT computer classroom. As a course reference, a textbook will be provided, The Art of Modeling with Spreadsheets: Management Science, Spreadsheet Engineering, and Modeling Craft. Upon the completion of the course, the participants will: 1. Recognize business problems that are amenable to established OR/MS methods rather than having to approach every problem ad-hoc. 2. Become better able to direct the work of their staff analysts. 3. Be knowledgeable of the toolbox of analytical software for a. Forecasting b. Decisions with multiple criteria c. Optimization d. Estimations with parameter uncertainties e. Deciding the best course of action given unpredictable future events 4. Be aware of the consulting resources available for advanced analysis. 5. Be better able to prescribe training for their analytical staff. Instructors: Dr. James Dyer, Fondren Centennial Chair in Business and Dr. Leon Lasdon, David Bruton Jr. Chair in Business Decision Support – both in the McCombs School of Business, The University of Texas at Austin. Course Topic Overview The course will begin with forecasting, which is accomplished with various statistical tools. For example, analysts commonly use the simple regression capability in Excel, but often have no formal training in its use and pitfalls; likewise for the more difficult tool of multiple regression. Also covered will be the other fundamental forecasting approach, time series analysis, the state of the art and science of projecting trends into the future. The second topic, closely related to the first, addresses estimates and forecasts when many underlying variables are uncertain. The “Monte Carlo simulation” method was designed to deal with this situation. It produces not only the needed estimate, but also information about how “solid” that estimate is, which can be very useful for decision makers. The third topic concerns the most common form of decision problem: choosing, prioritizing or allocating among alternatives given multiple criteria for evaluation. While this is frequently done ad-hoc on a spreadsheet, a process has been developed to formally structure the decision in a way that allows more precise capture of values and preferences, thus providing insight into the decision as well as a solution. The formalization is called multi-attribute utility theory, upon which numerous software applications are based. The fourth topic addresses optimization, defined as either 1) making decisions which result in the maximum benefit within a budget or 2) meeting a desired level of service at minimum cost. This session will employ the optimization technique of linear programming, which can quickly provide the best answer as opposed to trial-and-error, which could continue indefinitely and inconclusively. Improving Operational Decision Making With Spreadsheet-Based Modeling - Syllabus - Topic 1: Forecasting (Wednesday) Determining interrelationships and regression analysis; time-series forecasting methods; sensitivity analysis. Case study – forecasting TxDOT fuel-tax revenue Short-term [time series] Intermediate-term [simple regression] Long-term [multiple regression, lagged variables] Topic 2: Risk Analysis in Excel (Thursday morning) When risk-analysis simulation is needed; widely used Excel simulation add-ons; model formulation and interpretation of results Case study – applying @RISK software to TRENDS revenue forecasting model Topic 3: Multiple-criteria decision making (Thursday afternoon) Formulating a decision hierarchy; transforming criterion values to utility values; criterion importance weighting; scoring the alternatives; sensitivity analysis. Case study – selecting transportation research projects Case study – identifying dangerous intersections Topic 4: Optimization using the Microsoft Excel Solver (Friday) Spreadsheet engineering principles. Common types of optimization problems (capital budgeting, cash management, transportation, production and inventory management, goal programming); Excel Solver capabilities and limits; a framework for formulating optimization models, understanding their solutions, and communicating results. Case study – maintenance office location Case study – target setting and funds allocation to road tiers JAMES S. DYER The Fondren Centennial Chair in Business McCombs School of Business The University of Texas at Austin Professor Dyer received his B.A. with honors, Phi Beta Kappa, in physics in 1965 with minors in mathematics and philosophy from The University of Texas at Austin. He received his Ph.D. in business administration in 1969 with major field in statistics and minor fields in quantitative methods and management from The University of Texas at Austin. Dyer has been a professor at a major university since receiving his degree from graduate school in 1969. His academic experiences include the following: A member of the faculty of the McCombs School of Business since 1978, he served on the faculty of the University of California, Los Angeles for nine years, before returning to The University of Texas. In spring 1999 he was the Philip J. Rust Professor of Business at the Darden School of Business at the University of Virginia. He served as chair of the Department of Information, Risk, and Operations Management at The University of Texas for nine years (1988-97). This department offers courses in decision and risk analysis. He also served as chair of the Department of Management from 1982-1986. He teaches courses on risk analysis and investment science in the MBA and Ph.D. programs at The University of Texas. He also teaches in The University of Texas Executive MBA programs offered in Dallas and Mexico City. He has offered executive education programs in a number of cities in the United States, as well as in Australia, China, New Zealand, Singapore, Indonesia, Columbia, Ivory Coast, Mexico, Austria, England, Italy, Kuwait, Kazakhstan, and Angola. He served as chair of the Decision Analysis Special Interest Group of the Operations Research Society of America (now known as INFORMS). This group has a membership of over 1000 academics and practitioners in the fields of risk analysis and investment science. He has published three books and more than sixty articles on risk analysis and investment science. He received the College of Business Administration Foundation Advisory Council Award for Outstanding Research Contributions in 1999. LEON LASDON The David Bruton Jr. Chair in Business Decision Support McCombs School of Business The University of Texas at Austin Leon Lasdon received his Ph.D. in Systems Engineering from Case Institute of Technology in 1964. He taught in the Operations Research Department at Case Western Reserve University from 1964 to 1977, when he joined the Management Science and Information Systems Dept., McCombs School of Business, The University of Texas at Austin. He holds the David Bruton Jr. Chair in Business Decision support. Professor Lasdon currently teaches two MBA level courses: Decision Support Modeling and Financial Modeling and Optimization. He also teaches Ph.D. level courses in nonlinear optimization and large-scale systems optimization. He was one of three faculty named “outstanding graduate teacher” in 1984. In recent years he has incorporated an active learning approach into his classes, emphasizing student discussion and problem solving in small groups. Prof. Lasdon is co-author of the Microsoft Excel Solver. His GRG2 nonlinear optimizer is used in that solver. For more information see an article titled “Design and Use of the Microsoft Excel Solver,” which appeared in the INFORMS journal INTERFACES, Sept.-Oct., 1998. His research and consulting activities include optimization software, supply chain modeling and optimization, customer relationship management and other marketing issues, process control, and engineering design optimization. His nonlinear optimization software is widely used in financial asset allocation, option pricing, process control, and engineering design. He is the author or co-author of over 100 refereed journal articles and three books. Recent papers are available at www.utexas.edu/courses/lasdon (link to “papers”).