Heuristic optimization using computer simulation: a study of staffing levels in a pharmaceutical manufacturing laboratory

Abstract

This paper summarizes the results of the integration of four heuristic optimization techniques and a computer simulation model. In this project, our goal is to optimize the operating performance of a pharmaceutical manufacturing laboratory in which a small set of operators service a larger set of testing machines. The combinatorial complexity of the operator/test machine set along with the inherent nonlinearity, variability, and stochastic nature of pharmaceutical manufacturing make this problem very difficult, if not impossible to solve using traditional operations research tools. Thus, a very detailed computer simulation model of the laboratory was constructed to answer questions related to the capacity and cycle time capability of the laboratory. While this simulation model provided detail about the dynamics of the operation and functioned as a convenient what-if evaluator of proposed operational changes, it did little in terms of telling us what assignment of operators to test machines is best or what the potential range of improvement could be.