The Particle Swarm Optimization (PSO) algorithm is a kind of swarm intelligence(SI) optimization algorithm, which has been empirically shown to perform well on many optimisation problems. Because the principle of PSO algorithm is simple, and it is easy to implement, fewer parameters, converging to the global optimum with faster convergence speed, wide range of applications, etc., so PSO algorithm has received attention from the majority of research scholars since its introduced.Chaos is widely exists in nature, which is a kind of stochastic behaviors of determinate nonlinear system. Chaos has been applied in many disciplines successfully since its introduced. The research has made a lot of achievements in problem optimization, control systems, communications security, and other application fields.The basic PSO algorithm has some problems in solving complex function optimization, such as, falling in local optimum easily, losing of population diversity in later period of evolution, lower precision, etc. The basic principle and optimization process of PSO algorithm have been analyzed, chaos is introduced to the search process of PSO algorithm, two improved particle swarm optimization based on chaotic mapping are proposed in this paper as follows:(1) The simplified PSO algorithm based on controlled chaotic mapping for inertia weight is put forward, the structure of simplified PSO algorithm that discards the particle velocity. A set of chaotic variables is generated with classical chaos equation Logistic, adding control input to the set of variables, controlled chaotic state is mapped to the inertia weight. The mutation operation is executed if the stagnation steps are greater than the set value of individual extreme values and the global extreme in the evolutionary process. The experimental results verified that the new algorithm can significantly improve the convergence rate and precision of algorithm, and enhance the capability of escaping from the local optimal.(2) A new improved predator-prey PSO algorithm based on controlled chaotic mapping is proposed in this paper. Decreasing inertia weight adjusting strategy is adopted in order to gradually weaken the exclusion influence of predator particle on prey particle, which is described by controlled chaotic variables, adding White Gaussian Noise(WGN) to the current best particle according to the calculated probability value during running. By testing on benchmark functions, the experimental results show that the algorithm has done a lot of exploration works in the early stages during the run of the algorithm, over time, and the refining capacity on the optimal is increased gradually. The new algorithm has greatly improved the diversity of population and the accuracy of global optimal solution.Finally, the whole research achievements of this dissertation are summarized and further research directions are discussed. |