Spring 2009 ISyE Colloquium Series
University of Wisconsin-Madison
Department of Industrial & Systems Engineering
Michael C. Fu
Ralph J. Tyser Professor of Management Science in the Robert H. Smith School of Business, Institute for Systems Research, and Department of Electrical and Computer Engineering
University of Maryland, College Park
Michael C. Fu is Ralph J. Tyser Professor of Management Science in the Robert H. Smith School of Business, with a joint appointment in the Institute for Systems Research and affiliate faculty appointment in the Department of Electrical and Computer Engineering, all at the University of Maryland, College Park. His research interests include simulation optimization and applied probability, with applications in supply chain management and financial engineering. He has published (as co-author or co-editor) four books — Conditional Monte Carlo: Gradient Estimation and Optimization Applications (1997), Simulation-based Algorithms for Markov Decision Processes (2007); Perspectives in Operations Research (2006); and Advances in Mathematical Finance (2007) — and over 100 journal articles, book chapters, and conference proceedings. He served as Stochastic Models and Simulation Department Editor for Management Science 2006-2008 and Simulation Area Editor of Operations Research 2000-2005. His awards include the Allen J. Krowe Award for Teaching Excellence (1995), the University of Maryland Distinguished Scholar-Teacher (2004), the Institute for Systems Research Outstanding Systems Engineering Faculty Award (2002), the IIE Operations Research Division Award (1999), and the IIE Transactions Best Paper Award (1998). He is a Fellow of INFORMS and IEEE.
A Model Reference Adaptive Search Method for Global Optimization
(joint work with Jiaqiao Hu and Steven Marcus)
Model Reference Adaptive Search (MRAS) is a randomized search method for solving global optimization problems. The method works with a parameterized probabilistic model on the solution space and generates at each iteration a group of candidate solutions. These candidate solutions are then used to update the parameters associated with the probabilistic model in such a way that the future search will be biased toward the region containing high quality solutions. The parameter updating procedure in MRAS is guided by a sequence of implicit probabilistic models called reference models. We describe a particular algorithm instantiation of the MRAS method, where the sequence of reference models can be viewed as the generalized probability distribution models for estimation of distribution algorithms (EDAs) with proportional selection scheme. In addition, we show that the model reference framework can also be used to describe the recently proposed cross-entropy (CE) method for optimization and study its properties. We prove global convergence of the proposed algorithm in both continuous and combinatorial domains,and carry out numerical studies to illustrate the performance of the algorithm.
2:30–3:30 p.m., Friday, May 8, 2009
Room 1152 Mechanical Engineering Building