The Study Of Global Optimization Hybrid Intelligent Algorithm Based On Particle Swarm Optimization

Particle swarm optimization (PSO) is an adaptive stochastic optimization algorithm which based on the population search strategy. PSO has been widely used in neural networks, fuzzy system control, pattern recognition and so on for its simple concept and fast convergence. Therefore, it is significant to study and master the characteristics and rule of PSO in both theory and application areas. In addition, in view of its wide market prospect, it's also important in practice to extend its application scope.Basing on the analysis of standard PSO algorithm, this article systematically researches the application and improvement of PSO algorithm. The main work of this thesis is as follows:1 Two improved PSO algorithms are proposed: one is the PSO algorithm with the strategy of nonlinear decreasing inertia weight, the other is the PSO algorithm with the average information of swarm and keeping active strategy. The experimental results demonstrate that both of the two proposed algorithms are better than the standard PSO algorithm in the ability of global searching.2 In order to overcome the disadvantage of premature convergence and oscillation in later period, three kinds of hybrid particle swarm optimization algorithms with different mutation operators are proposed: the PSO algorithm with exponent decrease inertia weight and stochastic mutation, improved velocity of the PSO algorithm and adaptive mutation and the PSO algorithm with adaptive threshold mutation. The experimental results demonstrate that these three kinds of improved PSO algorithms are excellent in global searching.3 For constrained optimization problems, two kinds of hybrid particle swarm optimization algorithms are proposed: the improved particle swarm optimization algorithm based on outside point method for solving constrained optimization problems and the improved chaotic particle swarm algorithm for solving the nonlinear constrained optimization problems. Numerical experiments show that both of the two new algorithms are effective and robust.4 For solving zero-one nonlinear programming problems, we proposed a penalty function-PSO hybrid algorithm. It is shown in numerical experiments that this algorithm is simple and easy to implement with fast convergence and high accuracy.In general, the theory and application of PSO are analyzed comprehensively. Finally, the whole research contents are summarized, and further work is given.