具有工件相关学习效应的一般多机器流水车间调度问题研究

Research on general multi-machine flow-shop scheduling problem with job-dependent learning effect

针对目前具有学习效应调度的研究范围局限在单机或双机流水车间调度方面且学习模型过于简单的问题,建立了一般情况下具有工件相关学习因子、以最大完工时间为目标的多机流水线调度模型。在对模型有效求解方面,针对多项式算法和启发式算法的不足,提出引入智能算法进行求解的思想,将新颖的布谷鸟智能算法用于模型求解,设计了IMM编码用于编码转换,用An混沌映射进行种群初始和启发式算法随机替换策略以提高种群的质量和分散度,再结合迭代贪婪算法和Metropolis准则以提高局部搜索能力和避免早熟,建立了一种混合布谷鸟算法。仿真验证了该混合算法的有效性和优越性。

Due to the current research literatures of scheduling problem with learning effect only limited to single or double machines flow-shop scheduling and simple learning models,this paper proposed a multi-machine flow-shop scheduling model with job-dependent learning effect and makespan criterion.As for solving the proposed model,it analyzed the disadvantage of polynomial algorithms and heuristic algorithms,and presented a idea applied intelligent algorithm to the model.Then,it developed a novel hybrid cuckoo search algorithm by designing IMM code to transform continuous variables into job permutation,and using a chaos mapping initialize population and random replacement policy of heuristic algorithm to improve population quality and dispersion,and integrating iterative greedy algorithm and the Metropolis criterion to improve the local search.Simulation experiments show the effectiveness and superiority of the hybrid cuckoo algorithm.