Title: Some recent BO algorithms for stochastic simulations

Abstract: Bayesian optimization (BO) has been successfully applied in many areas and is now widely used in machine learning and simulation optimization.  In the optimization of stochastic simulation systems, some considerations include accounting for large (heterogeneous) noises, performance measures beyond the expectations and high dimensionality.  In this talk, we will introduce several BO algorithms we have developed recently to tackle several of these considerations, including the design of efficient budget allocation to better address the noise, and search schemes for the optimization of quantile/risk measures and high dimensional large-scale problems.  We will also describe the intuition behind these algorithms, their properties and performances.

Bio: Szu Hui Ng is an Associate Professor and Department Head for the Department of Industrial Systems Engineering and Management at the National University of Singapore.  She holds B.S., M.S. and PhD degrees in Industrial and Operations Engineering from the University of Michigan.  Her current research interests include simulation analysis and optimization, and its application in maritime transport and decarbonization.  She is a member of IEEE and INFORMS, and her email address is isensh@nus.edu.sg.

WED 05 OCTOBER 4:00 PM – 5:00 PM AEDT MELBOURNE

ZOOM MEETING ID: 873 1557 5255; PASSWORD: 778635