Hyper-heuristic For Distributed Heterogeneous Assembly Permutation Flow-shop Scheduling

In the past decades,traditional centralized production mode has become more and more unsuitable for modern manufacturing industry due to the trends of economic globalization and increasingly intensified market competition.A new distributed assembly production mode,which has the advantages of low cost,high flexibility,and anti-risk capacity,has attracted more and more attention in the manufacturing industry.Efficient scheduling of distributed assembly flow shop can not only shorten the production cycle and improve the production efficiency,but also improve the ability of manufacturing enterprises to cope with challenges such as diversification of product variety and customer needs.Thus,the research on the scheduling problem of the distributed assembly flow shop has important theoretical significance and application value.The researches on this problem mainly focus on homogeneous permutation flow shops.However,in a real manufacturing shop,the capability parameters of corresponding machines in the distributed factories are different due to inevitable machine deterioration and update,and therefore factories have the characteristic of heterogeneity.It is more realistic to study the distributed heterogeneous assembly permutation flow-shop scheduling problem(DHAPFSP).As a newly proposed algorithm conceptual model for solving complex optimization problems,hyperheuristic can solve problems across different domains and has become one of the research focuses in the field of computational intelligence.Thus,by studying the characteristics of DHAPFSP and its extended model,an hyper-heuristic is proposed to solve these problems.Detailed research contents are stated as follows:(1)A mathematical model of DHAPFSP is built with the objective of minimizing the maximum completion time and an estimation of distribution algorithm-based hyper-heuristic(EDA-HH)is proposed.Firstly,a search-stage-based encoding method is presented to both improve search efficiency and maintain potential solutions.Secondly,in the global search stage,the low-level heuristic(LLH)sequence obtained from EDA-HH operates on the population,aiming to obtain a high-quality and diversified product assembly sequence.Finally,a critical-path-based referenced local search strategy is proposed to search the job sequence in critical products in depth to further improve the optimization quality of the algorithm.Simulations and comparisons are both carried out on 540 benchmark instances and the results demonstrate the effectiveness and robustness of the proposed EDA-HH.(2)A mathematical model of the distributed heterogeneous assembly no-wait permutation flow-shop scheduling problem with sequence-dependent setup time(DHANWPFSP-SDST)is built with the objective of minimizing the maximum completion time and the EDA-HH is proposed for the problem.Based on EDA-HH,a problem-oriented LLH set is proposed so that the powerful exploration ability of the EDA-HH can be guaranteed.A mixed local search strategy is proposed to search the high-quality solutions from product sequence and job sequence to further improve the quality of superior population.The benefit of the presented mixed local search strategy lies in the excellent exploitation ability with substantially reduced computational cost.Finally,by further constructing the instance set for DHANWPFSP-SDST,and comparing it with the newly proposed typical algorithms,the experimental simulation results further verify the effectiveness and universality of the proposed EDA-HH.