基于改进蛙跳算法的分布式两阶段混合流水车间调度

针对考虑顺序相关准备时间的分布式两阶段混合流水车间调度问题,提出一种改进的蛙跳算法以同时最小化拖后工件数和最大完成时间.该算法通过启发式方法和随机方法对种群进行初始化,采取基于种群和记忆的种群划分方法,同时给出模因组质量评价方法,并根据模因组质量将所有模因组划分为最优模因组、最差模因组和其他模因组,每种类型的模因组分别采取不同的搜索策略,并分配不同的搜索次数,其中最优模因组不参与种群划分.选用一种多目标经典算法和两种近5年提出的算法作为对比算法,并与改进蛙跳算法的变体进行比较以验证模因组搜索新策略的有效性.通过对大量实例的计算实验结果表明,模因组搜索新策略有效,改进蛙跳算法能有效求解分布式两阶段混合流水车间调度问题.

An improved shuffled frog leaping algorithm for the distributed two-stage hybrid flow shop scheduling

To solve the distributed two-stage hybrid flow shop scheduling problem(DTHFSP) with sequence-dependent setup times, an improved shuffled frog leaping algorithm is proposed to minimize the number of tardy jobs and makespan simultaneously. The heuristic method and random method are adopted to initialize the population, population division is done by using population and memory, and then all memeplexes are divided into a best memeplex, a worst memeplex and other memeplexes according to the quality of memeplexes. Different search methods and the number of searches are applied to different kinds of memeplexes and the best memeplex is excluded from the population division. A multi-objective classical algorithm and two algorithms proposed in recent five years are selected as the comparative algorithms, and a variation of the improved shuffled frog leaping algorithm is used for comparison in order to verify the effectiveness of the new search strategy of memeplexes. The comparative analysis based on extensive instances shows that the new search strategy of memeplexes is effective and the proposed algorithm can solve the DTHFSP effectively.