基于改进型多目标粒子群算法的晶圆制造系统瓶颈工作站调度

Scheduling of bottleneck workstation in wafer fabrication systems based on improved multi-objective particle swarm optimization algorithm

为了尽可能提高瓶颈工作站利用率,在获得较高系统产能的同时得到一个合理的制造周期,构建了以最小化瓶颈工作站的平均加权提前/拖期时间和最小化瓶颈工作站流程时间为优化目标的改进型多目标粒子群算法,并对瓶颈工作站进行了性能分析.将准时交货和快速生产要求分别映射为瓶颈工作站平均加权提前/拖期时间和流程时间,并构建了多目标优化模型.通过改进速度和位置的更新机制,对陷入局部最优的粒子进行交叉操作,设计了用于瓶颈工作站调度的改进型多目标粒子群算法.在不同作业规模下从算法的稳定性、Pareto前沿质量、收敛速度及运行时间出发,进行了调度仿真试验.结果表明该算法对提高瓶颈工作站的调度性能是有效的、可行的.

To maximize the utilization of bottleneck workstation and achieve a high throughput with a rea sonable cycle time, an improved multiobjective particle swarm optimization algorithm （IMOPSO） was proposed with an objective function to minimize the total flow time and the weighed average time of earli ness and tardiness. The performance of the bottleneck workstation was analyzed. A multiobjective opti mization model was built based on the mapping requirements of rapid productions as total flow time and just in time delivery as weighed average time of earliness and tardiness of the bottleneck workstation. An IMOPSO algorithm was established for the scheduling of bottleneck workstation by improving an update mechanism of velocity and position with a crossover operation to the particle falling into local optimum. Based on stability, quality, convergence rate and run time, the simulation experiments were performed to evaluate the proposed algorithm under the conditions of different job scales. The results show that the pro posed algorithm is valid and feasible to improve the scheduling performance of bottleneck workstation.