Research And Application Of Adaptive And High-Dimensional Multi-Objective Differential Evolution Algorithm

Differential evolution algorithm is a swarm intelligence optimization algorithm based on evolutionary ideas and population differences.The principle is simple,the control parameters are few,and the global convergence speed is fast.The parameter setting strategy of the differential evolution algorithm and the combination of the differential evolution algorithm and other swarm intelligence optimization algorithms are important research contents of the differential evolution algorithm.Based on the previous research work,this article does some research on the method of adaptively adjusting parameters in the differential evolution algorithm and the high-dimensional multi-objective differential evolution algorithm.The main contents are as follows:1.According to the research history and current status of the differential evolution algorithm,its basic theoretical research and applied research are introduced;then introduced the basic principle,algorithm flow,main parameters and convergence analysis of differential evolution algorithm.2.In order to take into account global search ability and local optimization ability,a parametric adaptive differential evolutionary algorithm based on selective pressure is proposed.The basic idea is to construct a parameter adaptive adjustment method according to evolution algebra and individual fitness.In the early stage,the algorithm has a strong selection pressure,which strengthens the local optimization ability of the algorithm.In the later stage,the algorithm maintains a weak selection pressure to improve the global convergence.The standard Benchmark test function was used to test the performance of the algorithm and compared with the standard differential evolution algorithm to verify the effectiveness of the new algorithm.3.High-dimensional multi-objective optimization is a difficult point in multi-objective optimization,and target reduction is a more feasible and effective method for high-dimensional multi-objective optimization.This article proposes a cluster-based high-dimensional multi-objective differential evolution algorithm for target reduction.This new algorithm was organized based on a multi-objective differential evolution algorithm using elite selection and ranking strategies.According to the approximate pareto optimal front,all targets are clustered based on correlation distance,and then similar clusters are deleted to obtain a reduced target set.During the optimization process,the full target set and the reduced target set are searched periodically to take into account the convergence and operation efficiency of the algorithm.The high-dimensional DTLZ test function is used to test the performance of the new algorithm and compared with other algorithms.The results verify the convergence and effectiveness of the new algorithm.4.Mixed traffic flow signal control at urban road intersections is a complex multi-objective optimization problem in traffic control.Current research shows that evolutionary algorithms can solve such problems well.In this paper,a cluster-based multi-objective differential evolution algorithm for target reduction is used to simulate the mixed traffic flow signal control problem.The results show that the average vehicle delay average,vehicle stop average,traffic capacity,non-motor vehicle delay average and the equilibrium solution of pedestrian crossing waiting time validates the effectiveness of the algorithm.Figure[16]Table[8]Reference[60]...