Minimizing mean flow time on a single batch-processing machine for the burn-in test operation of semiconductor manufacturing

Abstract

This thesis considers a scheduling problem at a batch processing machine for the burn-in test operation of semiconductor manufacturing where a burn-in oven corresponds to a batch processing machine and the ready and processing times of jobs arriving at burn-in oven are deterministic and known. For the problem, the total flow time measure is considered to construct schedules in batch sequences so that each batch processing time is determined by one of the jobs in the associated batch that has the largest processing time. The objective is total flow times. A dynamic program algorithm is proposed to find the optimal solution (schedule) in a situation where jobs arrive in SPT order. For general situations, a heuristic method is exploited by first arranging weight factors for jobs for job sequencing and then adopting the dynamic program algorithm for batching decisions. Simulation experiment is performed for the efficient evaluation of the heuristic method.