The Research And Application Of Hybrid Shuffled Frog Leaping Algorithm In Limited Buffer Flow Shop Scheduling

In modern manufacturing enterprises,flow shop,as an important production mode,is a research hotspot of workshop scheduling.Compared with the traditional flow shop scheduling,the shop scheduling problem with limited buffer constraints is more in line with the actual production conditions of enterprises,and the research of using intelligent optimization algorithm to solve the problem of limited buffer also has a certain practical significance.As a kind of swarm intelligence algorithm,SFLA was found to have better performance in job shop scheduling than other kinds of optimization algorithms.Therefore,this paper proposed an improved hybrid shuffled frog leaping algorithm(HSFLA)based on the research background of limited buffer flow shop scheduling problem(LBFSSP),and applied it to the actual shop scheduling management system.The specific work of this paper is as follows.Firstly,the research status of LBFSSP and SFLA was analyzed and summarized,and the impact of limited buffer constraints on the maximum completion time(Makespan)was analyzed through experiments.Through the research on the algorithm principle and design idea of SFLA,the main defects and improvement direction of SFLA were given.The next step was to design HSFLA to minimize Makespan as the optimization goal to solve LBFSSP.In order to solve the problems that SFLA is easy to fall into local optimal and easy to produce illegal solutions when solving,this paper gives the improvement methods of each part of the algorithm.The innovation and improvement of HSFLA are reflected in the following aspects: 1)The population initialization mode of the original algorithm was improved by mixing some strategies in NEH,IG,VNS and other algorithms;2)An improved step length and illegal solution correction method based on adaptive mobile factor have been designed,and a new local search strategy has been proposed;3)In the process of population evolution,Gaussian mutation method was introduced,which effectively improved the weakness of the original algorithm easy to fall into local optimal.Then,a large number of simulation experiments were carried out by using the classic Taillard instances,and compared with the more advanced research results on similar problems in recent years,which verifies the superiority of the algorithm in this paper.The effect of buffer capacity on different scale problems was also analyzed through experiments.Finally,based on the above research results,with the production workshop of a wind power tower manufacturing plant as the background,a corresponding scheduling system was developed,and good application results were achieved.