A Number Of Improvement On PSO With Its Application

Optimization problems in engineering, technology, economics, management and scientific research in various fields such as the use of essential, the problem of solving its great concern to the people, to solve the objective function is to find the smallest or the biggest meet the conditions. Commonly used optimization method such as Newton method, conjugate gradient method, pattern search, simplex method, Roesnborkc law and PoweU problem is the solution at an initial point of the domain selected by iteration to find an extreme point.With the objective of human-depth understanding of the world, there are some optimization method in dealing with people facing complex issues, such as high-dimensional, multi-pole, the complex nature of such functions, the accuracy of the solution, or the time required to solve regard, the optimization of the results are unsatisfactory. This has made practical and effective optimization technique is very necessary. Commonly used methods such as the evolution of artificial neural networks, tabu search, simulated annealing, genetic algorithms and ant colony algorithm to solve the problem at the time, such as showing strong potential, which can be a reasonable period of time approaching the complexity of the optimal solution of the problem object. These algorithms involved in neuroscience, artificial intelligence, statistical mechanics, the concept of biological evolution, etc., are certainly a lot of natural phenomena as the basic structure of the algorithm, known as one of a number of intelligent optimization algorithm.More than ten years ago of the new optimization algorithm - Particle Swarm Optimization (PSO) has gradually become concerned about the direction of the research scholars as one. Because it is simple, fast convergence and less domain knowledge required for the characteristics of a wide range of concerns were. Even though the development of particle swarm optimization algorithm for nearly a decade, but both theory and practice have to be sophisticated. In this paper, the study of particle swarm optimization algorithm significance, and then with the introduction of PSO study some basic questions, including the basic concepts of optimization and classification methods. Subsequently, from the basic structure of PSO algorithm, the algorithm features to improve the methods of implementation models and application systems to do more research. The main research content of the following areas:PSO algorithm for the existing local extremum vulnerable, slow convergence and poor accuracy deficiencies, a method of a simplified, mainly for the characteristics of particle swarm optimization and the characteristics of the formula itself, the standard algorithm prone to premature convergence and poor performance characteristics of the global convergence, and some other method, often making changes in more complex algorithms, in order to avoid these problems. The use of a simplified idea, one type of function for the optimization problem can be simplified calculation, the use of simplified PSO with inertia weight corresponding to break through the classical algorithm of inertia weight range, through the simulation results achieved good results.Chaos as a wide-ranging nature of nonlinear phenomena, random, ergodicity, sensitivity to initial conditions, with stability and instability, the long-term behavior of the unpredictability of the characteristics of optimization problems for characteristics of chaotic series used to initialize the position and speed of particles, not to change the particle swarm optimization algorithm is initialized with the time of the random nature of the use of chaos and enhance the diversity of species and particle search ergodicity, at a large amount of the initial group on the basis of merit from the initial groups. At the same time iteration in the evolution, there are questions precocious algorithm can introduce a chaotic sequence of search algorithms can be generated in the iterative optimal solution of a number of local neighborhood, the inert particles to help escape from local minimum points and fast search to the optimal solution.PSO algorithm based on the existence of the back easily into local optimization, premature convergence problem appears, many studies have focused on the parameters of inertia weight on the improvement, because the value of a great help global search, fast convergence, but not easy to be precise solution; value in favor of small local search, can be more precise solution, but slow convergence, so the search according to the particles, and accordingly adjust the values of inertia weight, so a self-adaptive chaotic particle swarm optimizationThese improvements in the algorithm through simulation experiments prove that the proposed improvements in the effectiveness of the method.