Department of Industrial and Systems Engineering
National University of Singapore
Many practical decision problems naturally exhibit multiple conflicting performance criteria, and should be solved with consideration for each objective simultaneously. The goal of multiobjective optimization is to identify a set of best trade-off solutions, thus providing more flexibility for the final decision. In many applications, the objective functions are all unknown and can only be evaluated via expensive simulation experiments. For example, in the oil and gas industry, numerical reservoir simulation is used to optimize production over the full lifetime of the reservoir while maintaining a minimum level of production over a short time horizon. In this talk, we consider unconstrained black-box biobjective optimization problems in which analytic forms of the objective functions are not available, and function values can only be obtained through computationally expensive simulations. We propose a new algorithm to approximate the Pareto optimal solutions of such problems based on a trust-region approach. The algorithm is designed to find a diverse set of solutions that is evenly distributed on the Pareto front, in order to provide decision-makers with enough information to make the final trade-off between objectives. To do this, we first identify local optimal solutions in the current trust region while a new trust region is constructed around the most isolated point in order to explore areas that have not been visited. We prove convergence of the method under general regularity conditions, and present numerical results suggesting that the method efficiently generates well-distributed Pareto optimal solutions.
Dr. Sujin Kim is an assistant professor in the Department of Industrial and Systems Engineering at the National University of Singapore. She received her Ph.D. degree in Operations Research from Cornell University in 2006. Before she joined the National University of Singapore, she was a visiting assistant professor in the Department of Industrial Engineering at Purdue University. Her research concerns simulation methodology, stochastic simulation-based optimization, and black-box optimization with applications in energy and health service systems. Her research has been funded by Singapore Academic Research Fund, ExxonMobil Research & Engineering, and Singapore Civil Defence Force.