Optimisation via simulation: a metamodelling-based method and a case study

Simulation is an important tool to make design and operation decisions to enhance the performance of manufacturing processes and systems. The coordinated use of optimisation methods and simulation models can result in better decisions. This paper presents an optimisation via simulation (OvS) strategy based on design of experiments and metamodelling. The strategy is satisfactorily evaluated in terms of effectiveness and efficiency using a series of global optimisation test functions. Additionally, a full-scale application of the proposed strategy to increase the throughput in a painting line of a company that manufactures automotive parts is detailed and discussed. The solution to this case study compares favourably to the results using a commercial OvS package. [Received 03 February 2010; Revised 02 September 2011; Accepted 02 October 2011]