Research On The Improvement Of Particle Swarm Optimization For Different Modal Problems

With the improvement of the computational accuracy requirements of computation intelligence,the corresponding optimization algorithms have been improved continuously.Therefore,the solution to the optimization problem has also received the continuous attention and research of many scholars.The optimization problem represents that under a certain condition,the optimal solution set of the target problem is searched or calculated by a series of methods according a certain evaluation standard.The solution of optimization problem also has become an important research direction in the field of control.Particle swarm optimization is a meta-heurist algorithm,which has many advantages such as fewer parameters and faster computation speed.However,because of the shortcomings of the iteration mechanism and the methods of population updating,the ability to deal with a lot of high-dimensional complex target problems and problems with more no-linearity and optimization target uncertainly is greatly reduced.Therefore,this article starts from operation mechanism,population and particle renewal mode,and summarizes its shortcomings.Two different improved particle swarm optimization algorithms are proposed to deal with the single modal high-dimensional complex problem and multimodal optimization problems respectively.Finally,the performance advantage of algorithm in theory and application is verified by a series of experiment test.The main work is summarized as follow:Firstly,an algorithm based on Hybrid Students Distribution Quantum Particle Swarm Optimization(HTDQ-PSO)is proposed for single-modal problems with the complex high-dimensional and nonlinear characteristics.The updating mode of quantum behavior particles and the mutation mode of adaptive dynamic student distribution operator is introduced,which greatly improves population diversity and particle searching performance.The simulation results show that the improved algorithm has obvious advantages over other common algorithms in dealing with complex high-dimensional multi-peak problems.Secondly,in order to verify the optimization ability,the algorithm is applied to the nonlinear system identification research based on the interference of heavy tail noise.The experimental results show that the algorithm performs well in both identification error and performance index evaluation.Finally,in order to provide decision makers with more optimization information,an algorithm based on adaptive Species Evolution BFGS Particle Swarm Optimization(SEB-PSO)is proposed to solve the multi-modal optimization problem in the face of the overall optimization problem with many global extremum and local extremum interference.This method is improved from dynamic parameter updating,sub-population core selection,species gradient evolution and redundant particle initialization.The algorithm is proved by experiments that it has good performance in multi-modal optimization problems,where in terms of convergence precision and convergence speed or in the face of high-dimensional complex problems.