Improved Particle Swarm Algorithm And Its Applications

Particle Swarm Optimization(PSO)was proposed in 1995,and its emergence has attracted the attention and research of many scholars.This algorithm is different from traditional optimization algorithms.It is actually a new type of swarm intelligence optimization algorithm.The main advantages are its simple structure,low parameters,fast particle convergence and strong search capability.Since the advent of the algorithm,it has been widely used in many fields and has obtained better optimization results,including combinatorial optimization,neural network training,and data mining.However,this algorithm also has some drawbacks,such as the problems of premature convergence and low optimization accuracy.An improved PSO algorithm is proposed based on the problems of traditional algorithms.The purpose is to improve the convergence speed and global optimization ability of the algorithm.The proposed algorithm is applied to the photovoltaic power generation system and feature selection problems,so as to further verify the validity and value of the proposed algorithm.The main work is as follows:First of all,the principle and formula of optimization algorithm are systematically studied in this paper.Including particle swarm optimization,quantum-behaved particle swarm optimization and grey wolf optimizer.At the same time,improvement strategy of PSO algorithm is summarized based on PSO.After that,by comparing the experimental data,verify the reliability of the proposed algorithms.Secondly,one of the improved PSO,QPSO is a chaotic quantum-behaved particle swarm optimization algorithm which is easy to fall into the local optimum.Using the new chaos formula instead of the original evolutionary formula,it strengthens the random search property.Then,based on the improved algorithm,a new search mechanism is added,which enables the particle to quickly escape from the local optimal solution to seek the global optimal solution,thus effectively improving the global search ability and convergence speed of the algorithm.Finally,the suggested algorithm is used to the maximum power point tracking of photovoltaic power system.It can optimize the tracking efficiency of photovoltaicpower and effectively reduce resource consumption.Experiments have proved the effectiveness of the algorithm.Thirdly,based on thedisadvantage of low precision of PSO algorithm,this paper presents a new particle swarm optimization algorithm.The algorithm mainly adopts the basic idea of GWO algorithm and combines with the basic PSO algorithm to form a new PSO algorithm.In the evolutionary formula of the algorithm,the original empirical parameters are changed,and the initial search parameters are proposed to improve the population diversity and effectively solve the premature convergence of the algorithm.At the same time,the particle population is divided into hierarchical forms,and each particle is responsible for their search tasks.Sub-particle swarms move towards the optimal solution by elitist particles,which improves the accuracy of the algorithm and balances the global and local search ability of the particle.Finally,the algorithm is applied to the feature selection problem,in order to prove the validity of the proposed algorithm.In this paper,the basic test function is used to verify the feasibility of the proposed algorithm.The experimental results show that the basic PSO algorithm is better in the one-peak function,but the optimization ability of multimodal function is poor.However,in single peak and multi-peak functions,the group algorithm can achieve good results.