Improved Particle Swarm Optimization Algorithm

With the development of modern society, both in scientific research and engineeringapplications, there are a variety of optimization problems, such as portfolio investment, fi-nancial analysis, signal processing, fitting function. Therefore, optimization theory and algo-rithm is widely concerned and constantly exploring the subject. The optimization problemis nonlinear, nonsmooth, high dimensional, complex problem is not continuous, the tradi-tional derivative based algorithm in solving the optimization problem is subject to certainrestrictions, in order to solve these problems, we propose stochastic optimization methods,such as ant colony algorithm, artificial fish-swarm algorithm, artificial bee colony algorith-m, bacterial foraging algorithm, particle swarm optimization algorithm of swarm intelligenceoptimization algorithm. Swarm intelligence computation using bionics theory, by using com-puter reconstruction and Simulation of biological characteristics, such as biological habits,adaptive behavior, the compiler design optimization algorithm. Optimization of practicalproblems based on algorithm. Bionic algorithm optimization problem is now widely accept-ed and successfully solved already emerge in an endless stream, these algorithms becomesolving engineering, network optimization, an efective approach to complex issues such asintelligent control.Since the proposed particle swarm algorithm, the people proposed many improvedmethods.Clerc[1] proposed the shrinkage factorχ,By introducing a shrinkage factor to con-trol parametersω,c1,c2.Riget and Vesterstorm[2]proposed a kind of attracting difusion parti-cle swarm optimization algorithm.The algorithm introduces two operator”attract”and”difus-ion”,it improved the efciency of the algorithm.Mendes and Kennedy put forward FIP-SO[3],the limit position of particles change weighted combination for all particles.ZengJianchao and Wang Lifang proposed generalized particle swarm optimization algorithm, ex-tended the particle swarm algorithm.Monson and Seppi[4]proposed Kalman particle swarmalgorithm.Zhang Xuanping§Du Yuping§Qin Guoqiang proposed changed inertia weightadaptive particle swarm optimization algorithm[5].In addition,mixed strategies of diferen-tial evolution algorithm and particle swarm optimization,mixed strategies of chaotic searchand particle swarm algorithm can improve the algorithm abilities. 2011,Shi etc.proposed brain storming optimization algorithm[7],it is a new optimizationalgorithm inspired by traditional brain storming methods.The traditional brain storming col-lected all people.s ideas,but it was disturbed easily by other persons.s ideas.Therefore,brainstorming optimization algorithm introduced the group,it improved the algorithm.s perfor-mance.2012§Zhan etc.proposed modified brain storming optimization algorithm[8],it trans-formed K-means to SGM strategy.Then,it improved the convergent speed.Through carefully studying of the algorithms, we propose a modified particle swarmoptimization algorithm based on orthogonal design, it is an optimization algorithm by us-ing bionic optimization ideas and simulating human thinking mode.We used the modifiedparticle swarm optimization algorithm and other three bionic optimization algorithms to dothe experiments by typical test functions,learning from the chart,the modified particle swarmalgorithm had better stability and global search ability.Through the image observation,theconvergence speed of the modified algorithm is fast. Therefore, the algorithm had strongrobustness,high efciency and high precision of search.