Simulation-based Dynamic Scheduling Optimization For Automated Material Handling Systems In Semi-conductor

Semiconductor manufacturing is one of the most competitive industry in the world. Semiconductor enterprises must increase the on time delivery rate and decrease the cycle time to meet the dynamic demands. Automated material handling system (AMHS) plays an important role in 300mm wafer fab in terms of improving material handling performance and ergonomics. The performance of AMHS has a great impact on the performance of the whole semiconductor manufacturing system. This paper researches on the scheduling of AMHS.A simulation modeling framework is proposed in this paper, including input, simulation model and control and output. The simulation model is composed of releasing, manufacturing, material handling and scheduling module. The control integrates simulation optimization methods, including responses optimization and heuristic algorithm. A discrete event simulation model of intrabay system based on Arena was developed to implement the dispatching rulesA dynamic scheduling method is proposed, to accommodate the complex state of AMHS intrabay system. The method classifies the status of the system into three categories by critical factor and percent of waiting, i.e. urgent dispatching, long wait time dispatching and not long wait time dispatching. Multiple responses optimization method and simulation based Genetic Algorithm procedure are employed to search for optimized critical factor and percent of waiting as well as rules to apply under the three different situations. The result shows that the optimized dynamic scheduling rule outperforms longest wait time rule. Besides, simulation based Genetic Algorithm procedure is better than multiple responses optimization method in terms of performance and searching speed.In order to improve the performance of the whole system to some more extent, an integrated dynamic scheduling procedure is proposed. This procedure add releasing scheduling and lot scheduling to material handling scheduling, which classifies the status of the system into two categories respectively by stocker availability and queue length, i.e. high stocker availability and low stocker availability, long queue length and short queue length. Several rules are selected to apply under these circumstances. Simulation based Genetic Algorithm procedure are employed to search for the optimized integrated dynamic scheduling method. The simulation result shows that this method is more efficient than the static scheduling method and the material handling system dynamic scheduling method.The results imply that the scheduling optimization of releasing, lot scheduling and material handling can obtain better performance on on-time delivery rate, cycle time, work in process and wait time. Therefore, the dynamic scheduling method proposed can improve the performance of the whole system and it is useful in lowing the cost, insure the on-time delivery and increase the productivity.