论文部分内容阅读
针对机器随机故障下的柔性作业车间调度问题(Flexible Job Shop Scheduling,FJSP),提出了一种兼顾调度鲁棒性与稳定性指标的改进两阶段多种群遗传算法。算法采用基于工序和机器的双层编码方式,并根据机器故障概率插入空闲时间构成染色体;设计了基于非线性排序的轮盘赌法选择算子、改进的RPOX交叉算子和工序码机器码双变异的互换变异算子;在算法的第二阶段采用融合了进化代数、最大适应度和平均适应度信息的多种群自适应遗传算法,实现针对子目标和综合目标的分别进化,提高了算法的搜索效率保证了算法的收敛性。最后的仿真结果表明了该算法的有效性。
Aiming at the problem of flexible job shop scheduling (FJSP) under random failures of machines, an improved two-stage multi-population genetic algorithm with both robustness and stability of schedule is proposed. The algorithm adopts double-layer encoding based on process and machine, and inserts chromosomes according to the probability of machine failure into idle time. The roulette selection algorithm based on non-linear ordering is designed. The improved RPOX crossover operator and machine code double In the second phase of the algorithm, a multi-population adaptive genetic algorithm based on evolutionary algebra, maximum fitness and average fitness information is adopted to achieve the evolution of sub-goals and integrated goals respectively, and the algorithm is improved The search efficiency ensures the convergence of the algorithm. The final simulation results show the effectiveness of the algorithm.