Some efficient simulation budget allocation rules for simulation optimisation problems

In service industry, various decisions need to be made to design these service systems or improve their performances. In the face of complex systems and many choices, simulation is used to estimate the performance measures of each alternative when analytical expression is too complex or even unavailable. As multiple replications are required for each design, there is a need to efficiently allocate the simulation budget. The Optimal Computing Budget Allocation (OCBA) is an approach that intelligently allocates simulation budget for maximising the desired selection quality in finding the best alternative(s) and has demonstrated its ability in significantly enhancing simulation efficiency. In this paper, we present three latest developments on OCBA for the optimal subset selection, constrained optimisation, and multi‐objective optimisation problems. The models, the corresponding asymptotically optimal allocation rules, are provided together with numerical results showing their efficiency. The proposed rules are also further discussed from the large deviations perspective.