Improvement Of Multi-objective Evolutionary Algorithm And Its Application On Flexible Job Shop Scheduling

In scientific research and engineering practice,it often happens that multiple goals need to be optimized at the same time.As the problem dimensions increase,traditional optimization approach have poor performance.The evolutionary algorithm inspired by biological intelligence explores the solution space in the way of population evolution and has strong global optimization capabilities.It provides new ideas for solving multi-objective optimization problems(MOPs).For discrete manufacturing enterprises with flexible production methods of multiple varieties and small batches,in order to improve production efficiency and save production costs,it is necessary to design a reasonable workshop scheduling plan.The excellent performance of multi-objective evolutionary algorithm provides effective methods and means for solving such problems.This paper aims to improve the overall performance of the algorithm,and considers the convergence and distribution of the algorithm to improve the multi-objective optimization algorithm.And use multiobjective optimization theory to solve and optimize the flexible job shop scheduling problem(FJSP).The specific research contents are as follows:1.For the problem that the convergence and distribution of multi-objective particle swarm optimization are difficult to balance on MOPs,a vector angle-based competitive particle swarm optimization(Va CSO)is designed.Firstly,the algorithm performs clustering based on two indicators of convergence and distribution.Secondly,in order to tap the potential information of the particles,a competition mechanism based on vector angles is adopted to maximize the transmission of useful information in the particles.Finally,a population gap optimization strategy based on assisted learning is adopted to eliminate the problem of population gap in the target space caused by population clustering.The experimental results show that the comprehensive performance of Va CSO is significantly better than the comparison algorithms.2.In order to solve the problem of significant performance degradation of traditional multi-objective evolutionary algorithms in solving large-scale multiobjective optimization problems(LSMOPs),a decomposition-based large-scale three-particle competition algorithm(LTCSO/D)is designed.Firstly,to solve the problem of increasing the number of local optimums caused by large-scale,a population grouping based on spatial decomposition is proposed.Secondly,to solve the problem of insufficient diversity caused by the consistent search direction of particles,a three-particle competitive search strategy is proposed.Finally,in order to make full use of the competitive angle information,the environment selection based on vector angle is adopted.Comparative experiments show that LTCSO/D has significant effects on large-scale optimization problems.3.This article conducts optimization research for FJSP.Firstly,using the relevant theories of multi-objective optimization algorithms,a suitable mathematical model is established.Secondly,combined with the characteristics of discrete optimization,a multi-objective migratory bird group optimization algorithm(MOMBO/CF)based on cross fusion was designed.Finally,MOMBO/CF is used for the optimization of flexible job shop scheduling,and the effectiveness of the algorithm is verified through comparative experiments.Based on the experimental results,an effective scheduling plan with the maximum processing time and the total machine load as the optimization objectives is provided.