论文部分内容阅读
针对具有零等待约束的flow shop问题,以总流程时间和最大完工时间为多目标,提出一种结合多目标变邻域搜索的混合差分进化算法(multi-objective differential evolution hybridized with variable neighborhood search,M DEVNS)进行求解。提出一种基于改进Naw az-Enscore-Ham(NEH)规则的多样化种群初始化方法;设计了差分进化的变异、试验、目标个体更新操作;为提高多目标搜索能力,在算法的进化中混合了一种多目标变邻域搜索方法。通过Taillard标准测试算例的计算试验,证明了MDEVNS算法获得的Pareto前沿解在多样性和性能方面要优于多目标模拟退火算法和非支配排序遗传算法,验证了MDEVNS算法求解多目标零等待流水车间调度问题的有效性。
Aiming at the problem of flow shop with zero waiting constraints, a multi-objective differential evolution hybridized with variable neighborhood search (M) algorithm is proposed based on the total flow time and the maximum completion time. DEVNS) to solve. This paper proposes a method of population initialization based on the improved Nawaz-Enscore-Ham (NEH) rule. It also designs the mutation, test and target update of differential evolution. In order to improve the multi-target search ability, A Multiobjective Variable Neighborhood Search Method. Through the calculation of Taillard standard test case, it is proved that the Pareto front solution obtained by MDEVNS algorithm is superior to the multi-objective simulated annealing algorithm and non-dominated ranking genetic algorithm in terms of diversity and performance. It is verified that the MDEVNS algorithm solves the problem of multi- The effectiveness of the shop scheduling problem.