在考虑工人技能学习差异的基础上,为解决多工人协作柔性车间调度问题,提出了基于稀疏邻域带精英策略的快速非支配排序遗传算法(Non-dominated Sorting Genetic Algorithm Ⅱ,NSGA-Ⅱ)的调度方法.对考虑技能学习差异的多工人协作柔性车间调度问题进行了描述,以车间工人学习能力为背景改进了 DeJong学习模型,并建立了多工人协作柔性车间调度的多目标优化模型.在NSGA-Ⅱ基础上,引入了邻域稀疏度的选择方法,有效保留了信息丰富和多样化的染色体,并将稀疏邻域NSGA-Ⅱ应用于柔性车间调度问题求解.经实验验证,稀疏邻域NSGA-Ⅱ所得Pareto解集质量高于标准 NSGA-Ⅱ 和自适应多目标进化算法(Multiobjective Evolutionary Algorithm Based on Decomposition,MOEA/D),最短调度方案的完工时间为127.1 min,该方案满足逻辑和时间等约束.实验结果验证了稀疏邻域NSGA-Ⅱ在柔性车间调度中的优越性.
Abstract
On the basis of considering the difference of workers'skill learning ability,in order to solve the scheduling problem of multi-worker cooperative flexible workshop,a scheduling method based on sparse neighborhood Non-dominated Sorting Genetic Algorithm Ⅱ(NSGA-Ⅱ)was proposed.The multi-person cooperative flexible job shop scheduling problem under the premise of considering the difference of learning ability was described.The DeJong learning model was improved,and the multi-objective op-timization model of multi-person cooperative flexible job shop scheduling was established.On the basis of NSGA-Ⅱ algorithm,the selection method of neighborhood sparsity was introduced,which effectively retained the chromosomes with rich information and diversity,and then the sparse neighborhood NSGA-Ⅱ was applied to solve the scheduling problem.The experimental results show that the Pareto solution set quality of the sparse neighborhood NSGA-Ⅱ is higher than that of the standard NSGA-Ⅱ and the a-daptive Multiobjective Evolutionary Algorithm Based on Decomposition(MOEA/D),and the time of the shortest completion time scheduling scheme is 127.1 min,which meets the constraints of logic and time.The experimental results verify the superiority of sparse neighborhood NSGA-Ⅱ in flexible job shop scheduling.
维普
万方数据