Research On Flexible Job Shop Scheduling Problem Based On Intelligent Algorithm

In the context of economic globalization,the development of manufacturing cannot be ignored and must keep pace with the times.If manufacturing enterprises want to develop rapidly,they must meet customers’ demands for products intelligently and pluralistically,as well as customization and personalization of orders.Workshop production scheduling is a necessary link for enterprises to meet these conditions.Flexible job shop scheduling problem is more related to whether the enterprise’s production tasks can be completed scientifically and effectively in time,whether the utilization efficiency of equipment resources is efficient,and whether the enterprise’s production costs are lowest.In order to solve the problems of inefficient production planning,low utilization of equipment resources,and many uncertain factors in the process of job shop scheduling that affect the implementation of production planning,the production scheduling of job shop,especially the scheduling of flexible job shop,has been paid more and more attention by enterprises.In this thesis,the flexible job shop scheduling problem(FJSP)is taken as the research object.First,Genetic Algorithm(GA)and Non-Dominated Sorting Genetic Algorithm II(NSGA-II)are used to solve the traditional FJSP,and then the method is improved,Propose improvement methods to solve the changing needs of enterprises.The main contents of this thesis are as follows:1.First of all,for the traditional single objective FJSP,based on the traditional GA algorithm,this thesis introduces the Teaching-learning based optimization(TLBO)theory to improve the algorithm performance,and proposes an improved TGA to solve the problem.For the improvement of the inherent evolution mode of traditional GA,a new evolution mode is proposed based on TLBO theory.The overall combination of traditional crossover and teaching crossover is used to improve the solution efficiency of the improved algorithm.On this basis,parameter adaptation based on fitness value is introduced to further improve the solution performance of the algorithm through crossover and mutation adaptation.Relevant cases are cited for algorithm simulation solution,and the improved TGA result is superior to GA18.2% and HGBA10%.Its data shows that this TGA solution to FJSP has certain advantages.2.Based on the analysis of the current status of manufacturing enterprises and the impact of different production demands on scheduling,the FJSP with changing demands is proposed.In order to deal with this kind of problem,this thesis proposes an adjustable weight assignment algorithm to solve this problem.Combining traditional GA with NSGA-II,and combining the encoding and decoding methods of multi-objective algorithm,first determine the scheduling problem to be solved,use the fitness value to distinguish the advantages and disadvantages of different individuals,and then convert multiple scheduling goals into a single scheduling goal through the weight distribution method,and then solve it in the GA evolutionary way.Compared with traditional NSGA-II,the completion time is optimized by 15.73%,and the other indexes are basically the same,It can be concluded that its solution has local advantages.Then adjust the weight value according to the processing demand of the enterprise,solve the relevant examples,and compare the results with the average value solution and the traditional solution.It can be concluded that it can obtain a better solution on the demand target,which is more in line with the actual needs of the enterprise.3.On the basis of single objective FJSP,the traditional multi-objective FJSP is solved.On the basis of this problem,NSGA-II is used as the solution method.To solve the problem of insufficient accuracy of the traditional NSGA-II algorithm,different cross mutation combinations are used for the two-layer coding algorithm,improved elite cross strategy and improved insertion mutation are used as the evolution mode of the process chain,The improved crossover and directed mutation are used to update the evolution,and adaptive strategies are added.The elite retention strategy of the algorithm is adjusted to make it more suitable for the improved algorithm.Using a specific case,the improved NSGA-II algorithm results are compared with the traditional results and the previous results to verify the algorithm performance.Finally,the above research contents are summarized and conclusions are put forward.