Production planning for semiconductor manufacturing via simulation optimization

This paper is concerned with production planning in manufacturing, which can be loosely defined as the problem of finding a release plan for jobs that minimizes the total cost (or maximizes the total profit). Production planning is a challenging optimization problem due to the variability in manufacturing systems and uncertainty in future demand, both of which have not been adequately addressed by existing production planning models. To address both these issues, this paper formulates the production planning problem as a simulation-based multi-objective optimization problem, and adapts a genetic algorithm to search for a set of release plans that are near-Pareto optimal. The solutions from the simulation optimization approach can serve as a useful benchmark for existing and new production planning methods.