The Departmental Seminar Series is open to all. U-M Industrial and Operations Engineering graduate students and faculty are especially encouraged to attend.

Title:
Importance Sampling with Stochastic Computer Models: From Theory to Practice

Abstract:
Importance sampling has been widely used to improve the efficiency of deterministic computer simulations where the simulation output is uniquely determined, given a fixed input. To represent complex system behavior more realistically, however, stochastic computer models are gaining popularity. Unlike deterministic computer simulations, stochastic simulations produce different outputs even at the same input. This extra degree of stochasticity presents a challenge in analyzing engineering system performance. Our study tackles this challenge by addressing two problems. First, we derive the optimal importance sampling density and allocation procedure that minimize the variance of an estimator. Second, we present a non-parametric approach to approximate the optimal importance sampling density with a multivariate input vector when each factor’s contribution is different. The application of our method to a computationally intensive, aeroelastic wind turbine simulator demonstrates the benefits of the proposed approaches.

Bio:
Eunshin Byon is an Associate Professor in the Department of Industrial and Operations Engineering at the University of Michigan, Ann Arbor, USA. She received her Ph.D. degree in the Industrial and Systems Engineering from the Texas A&M University, College Station, USA in 2010. Dr. Byon’s research interests include data analytics, quality and reliability engineering, system informatics and uncertainty quantification. She has received several Best Paper Awards including the Best Applications Paper Award from IISE Transactions on Quality& Reliability Engineering. Dr. Byon has served the Quality, Statistics, and Reliability (QSR) subdivision of INFORMS as a chair-elect and chair in 2019-2020. She is a member of IIE, INFORMS and IEEE.