Simulation optimization with GA and OCBA for semiconductor back-end assembly scheduling

This paper presents a simulation optimization with genetic algorithm and optimal computing budget allocation for semiconductor back-end assembly scheduling problem to achieve minimal average order flow time. In particular, this research explores characteristics of hybrid flow shop scheduling problem in complicated identical and unrelated parallel machines, orders with specific product and demand scheduled with different release time, order split for parallel and merges for batch processing during manufacturing under product-machine dedication with stochastic processing times and sequence-dependent setup times. As this is a real-life stochastic event system and discrete simulation is usually the only resort for performance. Coupling genetic algorithms in simulation optimization as the solution space is large. Optimal computing budget allocation was used to reduce simulation budget usage while stochastic situation. Numerical result shows it is effective to improve performance better than practical heuristics and efficient to generate few replications which conquer the barriers utilize the advantages of simulation-based approach. This work can be as solution architecture for providing superior scheduling decision on allocation of orders and subsequent jobs to machines in a complex hybrid flow shop.