Optimal computing budget allocation for multi-objective simulation models

Simulation plays a vital role in identifying the best system design from among a set of competing designs. To improve simulation efficiency, ranking and selection techniques are often used to determine the number of simulation replications required so that a prespecified level of correct selection is guaranteed at a modest possible computational expense. As most real-life systems are multiobjective in nature, in this paper, we consider a multiobjective ranking and selection problem, where the system designs are evaluated in terms of more than one performance measure. We incorporate the concept of Pareto optimality into the ranking and selection scheme, and try to find all of the nondominated designs rather than a single "best" one. A simple sequential solution method is proposed to allocate the simulation replications. Computational results show that the proposed algorithm is efficient in terms of the total number of replications needed to find the Pareto set.