Improved Particle Swarm Optimization Algorithm

As a simplicity and efficiency meta heuristic evolutionary algorithm,Particle swarm optimization(PSO)can be used to solve complex problems in life,such as combinatorial optimization problems,nonlinear continuous optimization problems and mixed integer nonlinear optimization problems.And its practical application has attracted wide attention and recognization.In this paper,the defects of the particle swarm optimization algorithm is improved,especially the selection and control of parameters in PSO are studied.The main research contents are as follows:Firstly,a particle swarm optimization algorithm model based on improved inertia weight is established.Aiming at the problem of low accuracy and not easy to converge to the global optimum,the sine factor of random perturbation is added to the inertia weight.It has been proved by four test functions that the improved algorithm has higher precision and better fitting performance than standard particle swarm optimization algorithm.Secondly,Opposition-based particle swarm optimization based on Levy variation is established.In the initial stage of particle swarm search,use Opposition-based learning strategy to expand search space,and identify stagnant particles through location factor and speed factor in population updating,and then use these improved search strategies with Levy flight characteristics to update them.The simulation results show that the improved algorithm has a great improvement in the early search capability and the later search precision.Thirdly,aimed at the contradiction between the diversity and convergence of particle swarm optimization,an adaptive adjusted dynamic particle swarm optimization algorithm is proposed.The dynamic search space strategy is introduced into the particle swarm algorithm,and the adaptive adjustment of learning factor to balance the global search capability of the algorithm and the local search ability,simulation tests were carried out using four test functions to verify the effectiveness of the improved algorithm.Then three improved PSO algorithms are applied to solving the portfolio optimization problem,and the results show that three improved algorithms all can effectively reduce the risk of investment portfolio.Finally,the full text is summarized,and the further research problem of particle swarm optimization is prospected.