Abstract
This study presents a simulation optimization approach for a hybrid flow shop scheduling problem in an actual semiconductor back-end assembly facility. The complexity of the problem is determined on the basis of demand and supply. Demand varies with orders, which are in turn characterized by different quantities, product types, and release times. Supply varies with the number of flexible manufacturing routes; however, this factor is constrained in a multi-line, multi-stage production system that utilizes specific types and numbers of both identical and unrelated parallel machines. This study addresses a non-deterministic polynomial-time-hard and stochastic discrete optimization problem that is subject to numerous constraints, such as product—machine dedication and lot split and merge due to charateristics of production process. Die bond, wire bond and mold were considered as bottlenecks. Due to unbalance throughput during these three stage, there is a split behavior at die bond stage, then to wire bond for reducing flow time by parallel processing; thre is a batch behavior before mold, then batch processing at mold stage due to economic scale. In the latter, jobs that belong to the same order must be processed by the same machine type at each stage under stochastic processing and sequence-dependent setup times. A simulation optimization approach is developed in view of the complex and stochastic nature of the problem. The approach includes a simulation model for performance evaluation, an optimization strategy that applies either a genetic algorithm or particle swarm optimization, and a technique for acceleration via optimal computing budget allocation. Flow time is improved because of optimal assignment in terms of production line and machine type. Scenarios depicting the different levels of demand, product mix, and lot-split size are analyzed to reveal the advantages of the proposed simulation approach. Furthermore, lot split-size (limited 2 options) is included as a decision variable and is coupled with different meta-heuristics to enhance solution quality and practical heuristics. Future research directions are then recommended on the basis of the computational results.