Metaheuristic approaches for solving the scheduling problem in the semiconductor diffusion work area

Abstract. Motivated by a real case study of a semiconductor company, the paper addresses the scheduling problem in the diffusion area, which is recognized as one of the most critical stages in semiconductor manufacturing. This work area can be modeled as a two-stage hybrid flow shop scheduling problem with re-entrant flows. The first stage is the cleaning process, wherein impurities are removed from the wafer surface. The second one entails several tasks, such as diffusion, deposition and oxidation, to be performed into different furnaces. Based on the product recipe, wafer lots may undergo multiple operations in the second stage. The diffusion area also includes several production constraints, such as p-batching, machine eligibility and Q-time constraint, which increase the complexity of the problem. A novel metaheuristic named the Modified Harmony Search algorithm is proposed to minimize the average flow time of the system.