基于Petri网和GA的多目标动态优化调度问题研究

Petri Net and GA Based Approach for Dynamic Multi-objective JSP

以最小化生产周期、机床负载和机床与工人的费用为调度目标,基于受控Petri网和遗传算法提出了一种多目标作业车间调度方法,用于解决作业车间的加工受到机床、操作工人双资源制约条件下的动态优化调度。首先,给出了基于Parikh矢量的Petri网控制器简化设计方法,并应用该方法构建了基于紧急订单的受控Petri网模型;然后设计了基于Pareto的遗传算法,能够获得Pareto最优解集;尤其重要的是能够针对车间的动态扰动给出相应的调度方案。最后,通过实例验证了该方法的可行性和有效性。

A multi-objective scheduling method based on the controlled Petri net and GA is proposed to the dynamic job-shop scheduling problem（JSP） constrained by machines,workers.Function objectives of the proposed method are to minimize the completion time,the maximum load of machines and the total expense of machines and workers.Firstly,a controller designed method for Petri net based on Parikh vector is introduced,and the Petri net model is constructed for urgent jobs.Then,the genetic algorithm（GA） is applied based on the controlled Petri net model and Pareto.The set of Pareto optimum solutions can be obtained.Especially important,it is capable of generating alternative schedule after an uncertain disturbance takes place on a job shop.Finally,a scheduling example is applied to show that the proposed method is feasible and efficient.