Improvement And Application Research Of Particle Swarm Optimization Algorithm

Optimization is one of the important branches of computational mathematics,and many practical problems require an optimization method to solve.The optimization algorithm is divided into two categories:deterministic algorithm and randomness algorithm.The deterministic algorithm mainly includes gradient descent algorithm,conjugate gradient algorithm,branch and bound,Quasi Newton algorithm,etc.However,these optimization algorithms can not solve the large-scale nonlinear global optimization problem well.And the adaptability is also poor,for which people began to study the stochastic optimization algorithm.for example,genetic algorithm,particle swarm optimization,firefly algorithm,differential evolution algorithm.The particle swarm optimization algorithm is a typical stochastic optimization algorithm.Its idea from the fact that the birds live in nature,and the algorithm is easy to implement,the convergence speed is fast,and it is easy to find the global optimal solution.For the particle swarm optimization algorithm,this paper makes the following research:1.Introduce the background of optimization problem and its related definitions.On this basis,describes the background of particle swarm algorithm,algorithm idea and update formula,analyzes the characteristics and parameters of the algorithm and uses the dynamic system theory and Markov theory to prove the convergence of the particle swarm optimization algorithm.Particle swarm optimization does not guarantee global convergence.2.The paper proposes five improved particle swarm optimization algorithms models(1)There are three improved algorithms under the simplified particle swarm optimization algorithm model:IPSO1,IPSO2,and IPSO3.three improved particle swarm optimization algorithms use different local search strategy.(2)There are two improved algorithms under the exponential smoothing particle swarm optimization algorithm model,IPSO4 based on chaotic perturbation strategy and IPS05 based on twin clone selection strategy.(3)A hierarchical particle swarm optimization algorithm IPSO6 based on multi-group ideas is proposed.(4)A new particle optimization algorithm IPSO7 that originating from a deep search idea.(5)IMOPSO is proposed on the basis of the classical multi-objective particle swarm optimization algorithm.The update formula is added to the standard search direction,and the mutation and twin clone operation are introduced for the selection of pbest.These improved particle swarm algorithms perform well to some extent.The particle swarm optimization algorithm is applied to the optimization problems encountered in real life.(1)The improved particle swarm optimization algorithm is applied to a classical NP-complete problem-TSP problem.(2)Establishing a robot path planning model,using improved particle swarm optimization algorithm to solve robot path planning,(3)Supporting vector machine classification model,soft margin parameters and kernel function parameters using improved particle swarm optimization algorithm,(4)improved Multi-objective particle swarm optimization is applied to multi-objective portfolio problems.Compared with other intelligent optimization algorithms,the experimental results show that the improved particle swarm optimization algorithm is also outstanding in the application of optimization problems.