Solve The Complex Distributed Pipeline Scheduling Problem Based On The Distribution Estimation Algorithm

With the continuous development of society,manufacturing has entered into the globalization mode.Distributed manufacturing is an important research content of manufacturing,which has such characteristics as NP-hard,diversity,variability,multiple constraints,nonlinearity,dispersion,uncertainty,etc.In the distr ibuted manufacturing mode,the research work of the complex distributed flow shop scheduling problem mainly includes how the jobs are reasonably assigned to the factory,the processing sequence of the jobs in each factory,how to transport the processed jobs and how to assemble the completed jobs,which achieves optimal scheduling indicators.Researching complex distributed flow shop scheduling problems not only has important academic significance but also has certain application value,which has caused the attention of the scholars in the field of production scheduling.In view of the academic significance and application value of the complex distributed flow shop scheduling problem,the research and development of intelligent optimization algorithms based on this problem has been widely concerned by the theoretical community and the industry.Estimation of Distribution Algorithm(EDA)is a novel swarm intelligence evolutionary algorithm based on probability and statistics.It has obtained considerable resea rch results in the field of production scheduling.Therefore,this paper focuses on three kinds of important distributed flow shop scheduling problems based on EDA algorithm.The main work is described as follows:(1)For the distributed permutation flow shop scheduling problem with limited buffers,a Hybrid Estimation of Distribution Algorithm(HEDA)is designed to solve this problem.The optimization index is to minimize the maximum completion time.The first,on the basis of the lowest completion factory(LCF)rules,the against lowest completion factory mapping(ALCF)rule is proposed,and the mapping of each job’s order to problem solution is realized.Then,a local search based on Swap neighborhood and Insert neighborhood is designed,which further enhances the local search ability of HEDA.Finally,through the simulation experiments of several scale problems and comparison with other algorithms,the superiority of HEDA is verified.(2)Based on(1),the delivery center is further considered,and the distributed flow shop scheduling problem with limited buffers and delivery is proposed.Aiming at this problem,an improved Estimation Algorithm(IEDA)was designed to solve this problem.The optimization index was to minimize the maximum completion time.First,according to the characteristics of the problem,the earliest arrived at delivery center(EAD)rules and the converse earliest arrived at delivery center mapping(CEADM)rules are designed,and the positive and negative mapping of the job order of each factory to the solution of the problem is realized.At the same time,the vehicle loading job(VLJ)rules is designed to properly distribute the job to the loading vehicle.Then,a local search combining the first improved jump-out mechanism and the Insert neighborhood structure is designed,which to perform a detailed search of the high-quality solution regions obtained by the global search of the IEDA.Finally,the validity of IEDA is verified by the comparison of simulation experiments and algorithms of different problems.(3)Based on(1),further consider the job assembly,and propose distributed assembly permutation flow shop scheduling problem with limited buffers.Aiming at this problem,designed a Bayesian Statistical Inference-Based Estimation of Distribution Algorithm(BEDA).The optimization index is to minimize the maximum completion time.Firstly,a hybrid probability model is established that integrates the job position information matrix and the order relation matrix to effectively learn high-quality solution information and guide global search.Then,a local search based on product assembly is designed to further improve the search efficiency of the algorithm.The validity of BEDA was verified by simulation experiments and algorithm comparison tests of several test problems.