A New Improved Particle Swarm Algorithm

Particle Swarm Optimization algorithm derived from the movement behavior of bird populations, which in 1995 by the Kenndy and Eberhart, is a cluster based on the evolution of intelligent optimization algorithm technology. Because of its strong global search capability, less parameter settings, simple and easy achievement, and so one proposed, aroused the concern of many scholars, and has been developed rapidly, and was applied to various fields, such as the function optimization, neural network training, fuzzy control system, as well as other genetic algorithm applications.Simulated Annealing algorithm (SimulatedAnnealing SA) from the solid simulated annealing process, in 1982 by the introduction of combinatorial optimization Kirkpatrick, and other areas to solve large-scale composition, in particular the NP-complete combinatorial optimization problems effectively approximation algorithm. Metropolis acceptance criteria used, and with a group called the cooling schedule process parameter control algorithm, the algorithm in polynomial time, given an approximate optimal solution. SA algorithm is jumping out of the local extreme points the ability to find the optimal or near optimal overall, and the choice has nothing to do initial point. Simulated annealing algorithm because of its inherent characteristics intensive computing, the major problem is that running a long time and computer memory, and the integration with other algorithms to essentially reduce the computation time.This paper briefly reviews the theory of intelligent algorithm cluster the background. Summed up the three clusters intelligent algorithm - ant colony algorithm, PSO algorithm, artificial fish algorithm development of the basic theory and the status quo, then the particle swarm algorithm carried out a detailed analysis. Particle Swarm Optimization algorithm is a new bionic algorithm, introduced in the original PSO algorithm to the basic principles of mathematical description of the algorithm and process parameters on the basis of the standard is introduced PSO algorithm. Followed by 10 years of PSO improved algorithm research, improvement of the four ideas, and introduces several typical improved model. Finally, a new algorithm, simulated annealing will be thinking and mutation operator to introduce PSO algorithm, three different variations operator of the function, and with the standard particle swarm algorithm and a simple algorithm simulated annealing , the verification of the algorithm to avoid premature in a local optimum, and so there are fairly good results.