运用带有记忆库的遗传算法求解作业车间调度问题

Applying genetic algorithm with memory base to solve JSP

在遗传算法的基础上,提出了一种带有记忆库的遗传算法,用于求解生产调度问题。该算法通过轮换的方法,分析了记忆库充满后如何更新和识别相同个体的问题,从而达到将加工任务分配到不同的并行机器上去执行,以利于机器的负载平衡。仿真结果表明,运用带有记忆库的遗传算法不但使整个加工过程的执行时间得到优化,而且各类机器完成的操作数相同、使用的时间也较为平均,达到了设计目标。同时,该算法的计算速度较快,适用于较大规模作业车间调度问题的求解。

A modified genetic algorithm with memory base was proposed based on traditional genetic algorithms for solving Flexible Job--shop Scheduling Problems （FJSP）, which analyzed how to update the memory base when it was full and how to identify the same individuals through the rotation method, so as to allocate processing tasks to different parallel machines for execution. And this was beneficial to the machine＇s loading balance. The results of simulation showed that not only the whole processing time was optimized but also the number of operations by all kinds of machines was same and the spent time were average, as arrived at the designing objective. Meanwhile, this method was characterized by a high calculation speed, fitting for solving the large--scale job shop scheduling problems.