A genetic algorithm with feasibility-agnostic encoding and three-phase decoding for scheduling semiconductor manufacturing facilities under queue time limits

Abstract

Queue time (Q-time) limits, which refer to the allowable time intervals between consecutive processing steps, play an essential role in the semiconductor industry. This importance stems from the fact that violations of Q-time limits are highly likely to degrade product quality and yield, ultimately resulting in resource waste, reprocessing, and increased environmental impact. To address these challenges, we propose a novel genetic algorithm for solving scheduling problems in semiconductor manufacturing facilities with sequence-dependent setups, aiming to minimize a weighted sum of the makespan and the total time exceeding the Q-time limits. Specifically, we introduce a feasibility-agnostic encoding scheme and a three-phase decoding procedure that consists of job assignment, feasibility adjustment, and machine allocation. Furthermore, customized crossover and mutation operators are incorporated to enhance the exploration of the solution space. Comprehensive experiments on six datasets demonstrate the superiority of the proposed method compared to other metaheuristics. Based on the improvement of baseline metaheuristics when initialized with the solutions generated by the proposed encoding and decoding procedures, we verify the viability of the proposed method.