基于Petri网和GASA的JSP动态分类调度研究

针对车间作业的加工受到机床、操作工人等双资源制约条件下出现多种扰动的JSP调度问题,提出了基于受控Petri网和GASA对不同扰动进行分类处理的新方法。首先通过构建带有控制器的Petri网模型使系统的运行满足期望的目标,然后基于该模型把遗传算法和模拟退火算法相结合,以最小化最大完工时间为目标,基于机床故障修复时间、工人离岗时间及取消订单包含任务的多少进行分类调度,然后根据扰动恢复后剩余任务的多少决定是否进行再次调度,避免大范围调整造成的生产不稳定状态,最大限度的维持车间的生产能力。最后通过实例验证了算法的有效性和可行性。

Dynamic Classified JSP Scheduling Based on Petri Net and GASA

A new classified scheduling method based on the controlled Petri net and GASA was proposed to the job-shop scheduling problem (JSP) with multiple disturbances constrained by machines, workers. Firstly, a Petri net with controller was modeled to make the system running satisfy the anticipant performance; and then the hybrid genetic algorithm and simulated annealing algorithm (GASA) was applied based on the controlled Petri net model. Function objective of the proposed method was to minimize the completion time, and scheduling was classified based on machine repairing time, and worker leaving time, and task of order canceling. In order to avoid unsteady state of processing due to regulate in large scale, and the job shop production ability can be maintained farthest, it was decided whether it was rescheduled again based on remainder task after disturbance resuming. Scheduling example was employed to illustrate the effectiveness of the method.