Simulation-based Evolutionary Optimisation of Buffer Allocation Problem

In this work, we investigate single-objective optimisation of discrete-event simulation (DES) models with evolutionary algorithms. We study genetic algorithms (GAs) and perform comparative analysis of the results from DES models of serial production lines where average steady-state throughput is considered as the performance criterion. We perform empirical analysis of different GA settings to identify ideal parameters in simulation-based optimisation. We experiment on different sized problems. The empirical results have shown the efficiency of GAs in simulationbased optimisation of production lines through buffer allocation.