求解PFSP的集成多策略教学优化算法

Integrated Multi-Strategy Teaching-Learning-Based Optimization Algorithm for Solving PFSP

在标准教学优化算法的基础上,提出一种集成多策略教学优化算法(IMTLBO)用于求解置换流水车间调度问题(PFSP)。为了生成具有一定质量和多样性的精英种群,初始种群的20%使用NEHLJP1算法生成,其余个体使用反向学习法产生;此外,教学阶段采用基于惯性权重的分组教学、正弦TF策略和变邻域搜索,学习阶段采用双学习策略;最后,通过双局部搜索来提高算法精度。为检验其有效性,在三类基准实例上进行实验,结果表明了IMTLBO相比其它算法具有显著的寻优能力。除此之外,针对汽车连杆部件制造的大规模生产问题进行求解,大幅缩短了完工时间,进一步表明了IMTLBO求解PFSP的有效性。

Based on the standard teaching-learning-based optimization algorithm,an integrated multi-strategy teaching-learning-based optimization algorithm(IMTLBO)is proposed to solve the permutation flow-shop scheduling problem(PFSP).In order to generate elite population with certain quality and diversity,20%of the initial population is generated using NEHLJP1 algorithm,and the rest individuals are generated using reverse learning method.In addition,group teaching based on inertia weight,sine TF strategy and variable neighborhood search are adopted in the teaching stage,and dual learning strategy is adopted in the learning stage.Finally,the accuracy of the algorithm is improved by double local search.In order to test its effectiveness,experiments are carried out on three benchmark instances.The results show that IMTLBO has significant optimization ability compared with other algorithms.In addition,solving the large-scale production problem of automobile connecting rod parts manufacturing has greatly shortened the completion time,further demonstrating the effectiveness of IMTLBO in solving PFSP.