A Decentralized MILP Method for Rescheduling Semiconductor Assembly Systems with Re-Entrance and Time Window Constraints

In semiconductor manufacturing, a master schedule over several weeks' time horizon is typically created to optimize production. However, during the production run, the semiconductor assembly line is subject to random disruptions, such as machine failures, which may cause the master schedule to become sub-optimal due to delays or even infeasible by violating key time window constraints. In this paper, to handle such disruptions, a decentralized mixed integer programming approach is proposed to reschedule the system considering the re-entrant flows and time window constraints, in order to keep up with the original master schedule. The original master schedule is taken as a warm start, and the new adjusted schedule is obtained by the CPLEX Optimizer. In this way, the semiconductor assembly line can properly react to a machine failure without replacing the whole master schedule. The method ensures the products meet the time window constraints and the delay against the master schedule is minimized. The experimental results show that the proposed method can dramatically reduce the delay and time window violations caused by disruptions.