<title>Optimizing large complex systems by simulation: a practical approach</title>

Abstract

Simulation is the only general tool for analyzing performance of large complex systems. However, blindly using simulation to locate the optimal configuration of a complex systems is an impossible task for most real life problems because of two major difficulties: 1) Adequate accuracy of results requires long simulation time; 2) The number of possible configuration is too large for exhaustive search. In this paper, we will show that by appropriately using shorter simulation and applying ordinal optimization techniques we can more effectively resolve those issues. In the case study, we will apply our technique to a resource allocation problem in a manufacturing plant.