The Extended Forms Of Particle Swarm Optimization Algorithm And Their Applications

Particle Swarm Optimization (PSO) is a parallel searching algorithm with n +1 memory points. After analyzing and extending the PSO algorithm, this work made the following contributions:1 The fundamental recursive equation, procedure and structure of PSO were studied, the influence of parameter to efficiency, and topological structure of PSO were reviewed. Under some conditions, the track of particle movement was obtained and its convergence was discussed.2 Aimed at the nonlinear equations, a new approach which at first transforms the nonlinear equations into optimization problem, and then to solve the optimization problem by chaotic particle swarm optimization (CPSO) algorithm and another algorithm combining the particle swarm and method of complex optimization were proposed. The proposed algorithms were applied to model the relation of composite structures fatigue life with stress, temperature and moisture and to solve a nonlinear equations of geometric measurement of thin wall rectangle section.3 Considering the Traveling salesman problem (TSP), a method named MCPSO combining the method of complex (MC) and particle swarm optimization (PSO), was proposed to solve TSP by making use of the property of gradient and geometric divide point. Some particular operations were designed for the solution sequence of TSP, which makes the method MCPSO for continuation function can be used to solve TSP. Moreover, some improved optimization searches were proposed for the sequence of TSP, that provides a new idea for further study. MCPSO was applied to the problems such as the optimal moving path for printed circuit boards (PCB), and the results showed that it is feasible and effective. Another method which use the particle swarm optimization blockwise was also proposed, and the experimental results demonstrate the good efficiency of the algorithm in comparing with other conventional methods.4 Chaotic particle swarm optimization (CPSO) was proposed to solve dynamic optimization problems by introducing suitable tactics. Further, CPSO was appliedto optimize the feed-rate of Park-Ramirez bioreactor and a batch reactor with fixed boundary conditions and satisfactory results were achieved.5 The Pareto solution and weak Pareto solution were extensively employed in multi-objective optimization problems (MOP). However, Pareto solution and weak Pareto solution were a set, and it is still a problem to decide the final solution. Ideal Pareto solution (IPS) was proposed in this paper, and a method was designed to find IPS by adjusting particle optimum and global optimum of PSO, that makes PSO suitable to find IPS of multi-objective. The experimental results were given to show the effectiveness of the proposed algorithm.6 Since PSO was easy to be trapped into local minima in optimizing higher dimensional optimization problems, a novel method named conjugate direction particle swarm optimization (CDPSO), which combines conjugate direction method with PSO, was proposed. Experimental results on nonlinear parameter estimation showed the proposed tactics is efficient. Besides, based on the necessities in practice, the binary system form of particle swarm optimization (BPSO) was given.Finally, the whole research contents were summarized, and further researchdirections were indicated.