Using Clustering Analysis For Decision-Making With Multiple Stochastic Objectives

A number of researchers have successfully integrated stochastic computer simulation models with combinatorial optimization procedures that generate solutions for decision-makers. These integrated approaches often use nature inspired search heuristics that also possess a stochastic feature of their own. These integrated simulation optimization approaches have been primarily designed to address single objective optimization problems. Only a few approaches have been designed for multiobjective optimization where they generate a finite set of Pareto optima. This Pareto optimal set often contains a very large number of solutions, which could be overwhelming to the decision-maker. In this paper, an innovative approach that effectively reduces the number of the solutions while considering the stochastic nature of the objective functions is proposed. A detailed description of the proposed approach and a numerical example that demonstrates the performance are provided