Meetings and Events 2010
Past Meetings and Events, 2010
Micro Motives and Macro Behavior Redux: Brains, Choice, Simulation, and Systemic Stability by Margaret M. Polski, Ph.D.
Date: Wednesday, November 17, 2010 Nobel laureate Tom Schelling observed more than 30 years ago that individual decisions and actions can lead to significant unintended consequences at the system level. Recent examples of U.S. policymakers’ need to better understand the unintended consequences of micro behavior at macro levels abound: The crisis in the financial system; counter terrorism and counter insurgency operations; stabilization operations in Pakistan and Haiti. While the influence of micro motives on macro behavior is more widely appreciated today by both researchers and policy makers, it is rarely studied systematically particularly in complex social systems, which are characterized by their propensity for delivering strategic surprise. Simulation tools like agent-based modeling can help us explore these phenomena. However, evidence from research in the social neurosciences suggests that we cannot change behavior without changing brains: How can we integrate advances in behavioral economics, neuroscience, and computational science to better understand and reduce the impact of behavior in complex social systems? Dr. Polski will present an alternative model of decision making and facilitate discussion of the implications for developing simulation tools. Margaret Polski, affiliate research fellow at the Krasnow Institute for Advanced Study and affiliate scholar at George Mason University's Mercatus Center Financial Markets Working Group, is a political economist with over fifteen years experience leading strategy and transformation initiatives in business, government, and civic sectors. Her research interests include security, growth, regulation, and the science of human behavior. She is currently a Sr. Advisor with ASE/Booz Allen Hamilton and she maintains research affiliations with the Krasnow Institute for Advanced Study and the Mercatus Center at George Mason University. Dr. Polski has previously served as a senior advisor in the Africa Bureau of the U.S. Agency for International Development; a management consultant with the Diel Group and A.T. Kearney; and held executive positions with AMERICORD, Inc., Check Technology Corporation, Santa Cruz Imports, and BMC Industries. She has a Ph.D. from Indiana University, an M.P.A. from the Kennedy School of Government at Harvard University, and a B.E.S. from the University of Minnesota. |
Pathologies of System Dynamics Models or “Why I am Not a System Dynamicist” by Dr. Robert Axtell
Date: Thursday, October 21, 2010 So-called system dynamics (SD) models are typically interpreted as a summary or aggregate representation of a dynamical system composed of a large number of interacting entities. The high dimensional microscopic system is abstracted-notionally if not mathematically-into a ‘compressed’ form, yielding the SD model. In order to be useful, the reduced form representation must have some fidelity to the original dynamical system that describes the phenomena under study. In this talk I demonstrate formally that even so-called perfectly aggregated SD models will in general display a host of pathologies that are a direct consequence of the aggregation process. Specifically, an SD model can exhibit spurious equilibria, false stability properties, modified sensitivity structure, corrupted bifurcation behavior, and anomalous statistical features, all with respect to the underlying microscopic system. Furthermore, perfect aggregation of a microscopic system into a SD representation will generally be either not possible or not unique. Finally, imperfectly aggregated SD models-surely the norm-can possess still other troublesome features. From these purely mathematical results I conclude that there is a definite sense in which even the best SD models are at least potentially problematical, if not outright mischaracterizations of the systems they purport to describe. Such models may have little practical value in decision support environments, and their use in formulating policy may even be harmful if their inadequacies are insufficiently understood. Dr. Axtell is the Chair of the Computational Social Science Department of the Krasnow Institute for Advanced Study at George Mason University. He earned an interdisciplinary Ph.D. at Carnegie Mellon University, where he studied computing, social science, and public policy. His teaching and research involves computational and mathematical modeling of social and economic processes. Specifically, he works at the intersection of multi-agent systems computer science and the social sciences, building so-called agent-based models of a variety of market and non-market phenomena. His work has been published in Science, the Proceedings of the National Academy of Sciences, and been reprised in Nature, as well as appearing in leading field journals. His research has been supported by private foundations and governmental organizations. Stories about his research have appeared in many major magazines and newspapers. He is co-author of “Growing Artificial Societies: Social Science from the Bottom Up” (MIT Press). Click here to download the presentation slides from this meeting. |
“A Manpower Planning Model for the Military Sealift Fleet Support Command” by Dan Steeples
Date: Tuesday, September 21, 2010 The US Navy's Military Sealift Command (MSC) operates approximately 45 ships that provide the following services to Navy warships: (i) underway replenishment, (ii) towing services, and (iii) special missions. The unique feature about these ships is that they are crewed by Federal civil service mariners. The Military Sealift Fleet Support Command (MSFSC) is responsible for ensuring that the right number of mariners is employed in each job category in each year so as to staff these 45 ships at appropriate levels. Importantly, it is the mariner himself or herself, not the Navy, who decides the length of time spent in each shipboard assignment. There are two fundamental management issues facing the MSFSC: (i) how many civilian mariners should be employed in each distinct job category in each year of the planning horizon, and (ii) given that the number of ships and job skill requirements change over the years, how does MSFSC develop a hiring and promotion plan to ensure that the necessary requirements are met at minimum cost. The problem is formulated, and solved, as a mixed integer program having approximately 8000 variables and 2000 constraints. Dan Steeples is a Senior Research Analyst in the Resource Analysis Division at The Center for Naval Analysis. He holds masters degrees in statistics and operations research from The University of Kansas and The University of California at Berkeley, respectively. He has served as an adjunct faculty member at several universities around the country including, George Washington University, the University of Maryland, and George Mason University. Click here to download the presentation slides from this meeting. |
“Disjunctive Mapping: Changing The Way We Understand and Predict Customer Behavior” by Dr. Warren Lieberman and Dr. Michael Raskin
Date: Wednesday, July 28, 2010 Relative to the traditional statistical techniques that we have come to rely on, this presentation presents a fundamentally different way to analyze and predict customer behavior. In addition, new analytical tools are described that highlight where and how opportunities exist to modify customer behavior to better achieve desired outcomes. Potential application areas include forecasting demand at alternative prices under multiple competitive responses. In addition, this approach facilitates estimating the likelihood with which alternative competitive responses will occur. Many commonly used techniques to understand and predict consumer behavior presume an underlying functional relationship - a model - buried in confusing data. We take the position that these models are generally not good representations of human behavior. Furthermore, with desktop computing having become so powerful, it is now practical to challenge whether the modeling approaches that we have come to rely on represent the best paradigm for understanding and predicting consumer behavior. Underlying our approach, termed Disjunctive Mapping (DM), is the notion that there are generally multiple routes (sets of influences and decisions) leading to any outcome and their effects can be measured in terms of change in an outcome's probabilities. Rather than attempt to capture central tendencies or capitalize on dominant patterns, DM obtains its power by focusing on the multiple ways events occur. DM metrics enable users to measure the change in probability of an outcome due to the influence of any factor or set of factors in the data, without building models. A structured inquiry process allows it to offer direct, accessible, comprehensive, and prioritized measures in answer to practical questions. Dr. Warren H. Lieberman is President of Veritec Solutions. Veritec is a consulting and software development firm focused on helping companies optimize their pricing and revenue management (P&RM) capabilities. Veritec is based in Belmont, California with a staff office in Boston, Massachusetts. Dr. Lieberman serves on the Board of Directors for the Institute for Operations Research and the Management Sciences (INFORMS) as Vice President, Information Technology. INFORMS is a professional society for Operations Researchers with over 10,000 members. He recently served as Chair of the Revenue Management and Pricing Section of INFORMS and for three years was the Committee Chair for the INFORMS Revenue Management and Pricing Section Prize, awarded for the best contribution to the science of pricing and revenue management published in English. He is currently serving on the editorial board for the Journal of Revenue and Pricing Management. Dr. Lieberman previously has served as Chairman of the Yield Management Study Group of the Airline Group of the International Federation of Operations Research Societies (AGIFORS). He is an internationally recognized expert in revenue management, especially in ‘non-traditional’ implementations. By combining his Operations Research-based modeling skills with a strong emphasis on ensuring that P&RM programs reflect industry operations as well as business processes and policies, Warren helps clients in many industries. He pioneered the application of revenue management techniques in the cruise, timeshare exchange, and equipment leasing industries, providing both design and technical leadership. In the 1980's, Warren designed some of the earliest multi-method adaptive demand forecasting techniques used in revenue management systems. Dr. Lieberman's focus areas include performance measurement, quick wins that enable companies to obtain early gains from P&RM programs, education and training programs, and using multiple signals to support pricing and inventory control decisions. Dr. Lieberman received the B.S. degree in Mathematics with a specialization in computer science from the State University of New York at Binghamton. He holds a Ph.D. in Operations Research from Yale University. Click here to download the presentation slides from this meeting. |
“Monster Mash-Ups for Fun and Profit” by Steve Mack
Date: Thursday, April 29, 2010 OR is a science of numbers. But clients like explanations with pictures. Glitz sells. The ability of clients to fully appreciate the results of an analytic model can be significantly enhanced through visualization. Analytic and visualization software have reached levels of maturity that enable rapid prototyping through clever mash-ups. Solutions can be analytically sophisticated with intuitive interfaces and slick visualization to generate the desired “Oh wow!” from the person who writes the check. Steve will present a few example mash-ups and show you how it can be done. The techniques are relatively simple to employ and the models are fun to build. So hopefully you may walk out of the presentation with some modeling ideas that you can actually try yourself. Steve Mack has degrees in Chemistry and Operations Research. He has worked in OR for over 20 years. His professional interests include, quantitative modeling, mathematical programming, decision management and the holistic integration of analytics with qualitative judgment. He is also a trained facilitator who nevertheless enjoyed Chemistry because the molecules were rarely petulant. |
“Reading the Behavior Signature: Predicting Leader Personality from Individual and Group Actions” by Dr. Paul Sticha and Elise Weaver
Date: Thursday, April 8, 2010 The personality of a leader can be used to predict that leader’s actions as well as those of the group that he or she leads. However, except for a small number of well-known leaders, the personality of leaders must be inferred from actions and other evidence. We have developed a Bayesian network to infer leader personality variables related to violence from evidence of leader and group actions and the situational demands and context in which the actions occur. The network was applied to a historical situation, and its ability to distinguish extreme personalities was established. We describe the process used to relate a leader's behavior to his or her personality, based on the judgments of leadership researchers. We illustrate how the relationships can be used to predict the likelihood of leader actions, given personality and situational variables. Finally, we show how the model can be used to infer a leader's personality traits, given actions in context. Paul J. Sticha manages the Modeling and Simulation Program at HumRRO. He has conducted research that incorporates personality variables into Bayesian networks to use in models that predict leader behavior. Dr. Sticha’s work includes the development and validation of mathematical models of human information processing, the application of probabilistic and decision-analytic methods to aid individual and group problem solving, the development of computer decision support systems for planning and design, and the application of advanced technology to aid training and training management. Principal areas of competence are decision and cost/benefit analysis, Bayesian networks, decision support systems, human performance modeling, mathematical modeling, military recruitment research, project management, simulation research, skill acquisition and retention. He has a doctorate in Mathematical Psychology from The University of Michigan. Elise Weaver is a Senior Scientist in HumRRO’s Modeling and Simulation Program. Working with researchers from Carnegie Mellon University’s CERT program, she participated in the development of the Management and Education of the Risk of Insider Threat (MERIT) model. Her areas of social science research expertise include social and cognitive psychology, especially judgment and decision making. Her areas of technical expertise include the development of models and simulations using system dynamics, Bayesian networks, and neural networks, the elicitation and analysis of expert judgments using judgment analysis, the design of experiments, and statistical analysis. She has a Ph.D. in Psychology from Duke University. Click here to download the presentation slides from this meeting. |
Capital Science 2010, Presented by the Washington Academy of Science and its Affiliated Societies
Date: Saturday, March 27 - Sunday, March 28, 2010 On Saturday and Sunday, March 27-28 2010, The Washington Academy of Sciences (WAS) and its Affiliated Societies (including WINFORMS and IIE) will hold the fourth in the series of biennial pan-Affiliate Conferences, Capital Science 2010 (CapSci 2010). It will be held in the Conference Facility of the National Science Foundation in Arlington, VA at the Ballston Metro stop. The Conference will serve as an umbrella for scientific presentations, seminars, tutorials, and talks. These pan-Affiliate Conferences underline the fact that the Washington, DC area is not only the political capital of the country but, in many respects, the nation's intellectual capital – with several major universities and government laboratories that are the homes of an astonishing number of Nobel laureates. |
“Portfolio Decision Analysis” by Gregory S. Parnell
Date: Tuesday, March 2, 2010 We provide an introduction to portfolio decision analysis using examples from defense, intelligence, and environmental domains. Many resource allocation and R&D problems in these domains are similar to commercial or other public domains. We compare defense portfolio analysis with two commercial applications. We present and discuss some reasons why organizations do not use portfolio decision analysis. We offer four reasons for using portfolio decision analysis: the complexity of portfolio decision making, to create value for the organization, to provide stakeholders confidence that their needs are being considered, and to provide an objective and transparent decision rationale to give decision makers and overseers confidence that public funds are being well spent. We review portfolio decision analysis applications to illustrate best practices including value for resources, optimization subject to constraints, and Monte Carlo techniques. Dr. Gregory S. Parnell is a Professor of Systems Engineering at the United States Military Academy at West Point and teaches decision and risk analysis, systems engineering, and operations research. His research focuses on decision analysis, risk analysis, and resource allocation for defense, intelligence, homeland security, and environmental applications. He is also the Chairman of the Board and a senior principal with Innovative Decisions Inc., a leading decision and risk analysis firm. He is a member of the Chief Technology Officer and Information Assurance Panels of the National Security Agency Advisory Board and is a former chair of a National Research Council study on bioterrorism risk analysis. He co-edited Decision Making for Systems Engineering and Management, Wiley Series in Systems Engineering, Wiley & Sons Inc., 2008, and has published over 100 papers and book chapters. Dr. Parnell has served as President of both the Decision Analysis Society of the Institute for Operations Research and Management Science (INFORMS) and the Military Operations Research Society (MORS). He is a fellow of the MORS, INFORMS, and the International Committee for Systems Engineering (INCOSE). Dr. Parnell received his Ph.D. from Stanford University. Click here to download the presentation slides from this meeting. |
