A robust simulation optimization algorithm using kriging and particle swarm optimization: Application to surgery room optimization

Abstract

Simulation optimization is an endeavor to determine the best combination of inputs that result in the best system performance criterion without evaluating all possible combinations. Since simulation optimization applies to many problems, it is extensively studied in the literature with different methods. However, most of these methods ignore the uncertainty of the systems’ parameters, which may lead to a solution that is not robustly optimal in reality. Motivated by this uncertainty, we propose a robust simulation optimization algorithm that follows the well-known Taguchi standpoint but replaces its statistical technique with a minimax method based on the kriging (Gaussian process) metamodel. The particle swarm optimization algorithm is applied in consideration of the uncertainty of the problem parameters to solve for a robust optimal answer. Finally, the performance of the proposed algorithm is highlighted through examples of numerical simulation and comparisons with other methods in this area, and it is applied to the surgery room optimization problem as a case study.