Multi-objective Particle Swarm Optimization Of Job-shop Scheduling Problems With Multiple-strategy Combination

Under the background of information age,manufacturing industry is developing towards intellectualization,collaboration,globalization and clustering.One of the problems that restrict the development of manufacturing industry is how to realize production efficiently and maximize the benefits of enterprises.Therefore,for the managers of enterprises,a convenient,efficient and reasonable method is needed to guide the processing and scheduling of enterprises.In this paper,a typical Job-Shop scheduling problem(JSP)in manufacturing is studied in depth,and a method of combining MS-MOPSO algorithm with QUEST logistics simulation software based on multiple strategies is proposed,which is used to solve the production and processing optimization problem of parts in a practical factory.For the multi-objective job shop scheduling problem to be studied,the main factors affecting the whole process in actual production are comprehensively considered,and a multi-objective job shop scheduling model is established,which takes the maximum completion time,total tardiness time and total machine idle time as the optimization objective function.Combining with the existing research trends,a co-evolutionary population search model based on particle swarm optimization algorithm and grey wolf algorithm is proposed,and a leader individual selection mechanism based on average Frechet distance curve similarity matching and Pareto dominance is proposed to guide the evolution of the group,finally,the variable neighborhood search method with three neighborhood structures is used to enhance the local search ability of the algorithm.In order to qualitatively analyze the optimization ability of the MS-MOPSO algorithm,the commonly used multi-objective test case DTLZ(1-7)is used to compare the MS-MOPSO algorithm with the other three algorithms,and the results prove the effectiveness of the optimization.For the premature convergence of the algorithm in some test cases,perturbation mechanism is proposed to avoid the algorithm falling into local optimum.In order to apply the algorithm to the optimization of the actual discrete JSP problem,the ROV criterion is used to change the coding method of the algorithm,and the local search method of the improved neighborhood structure is adopted according to the coding characteristics of the JSP problem.Finally,several benchmark test examples of JSP are used to demonstrate the effectiveness of the algorithm in JSP optimization applications.In solving practical problems in processing factories scheduling order to fully fit the actual production,cost-effective and can be obtained corresponding processing program.In this paper,the MS-MOPSO algorithm is first used to optimize the specific processing problems of a certain factory,so as to efficiently obtain an effective Pareto solution set,and then use QUEST simulation logistics software to establish a corresponding three-dimensional environment for the processing flow,and the schemes obtained by the algorithm are simulated independently in software.The final processing plan is selected based on the criteria for evaluating the machine utilization index commonly used in processing.The combination of the fast features of the intelligent algorithm and the high degree of fitting of the logistics simulation software to achieve comprehensive optimization of the actual actual processing scheduling can effectively improve production efficiency and economy.