A genetic algorithm to solve master planning problems in semiconductor manufacturing

Abstract

We present a master planning problem that arises in semiconductor manufacturing. The problem consists in determining appropriate wafer quantities for several products, facilities, and periods of time. Different demand types, i.e. confirmed orders and forecasts, are considered. We use a combined objective function that takes production costs for in-house locations, subcontracting costs, inventory costs, and costs due to unmet demand into account. Demand fulfillments and capacity constraints are considered. We present a genetic algorithm to solve the master planning problem heuristically. The results of some computational experiments are presented. We compare the results for small size test instances with results obtained from a commercial MIP solver. The genetic algorithm is able to produce solutions with reasonable quality in little time.