Parametric Analysis Of Particle Swarm Optimization Based On Genetic Algorithm

Particle swarm optimization, abbreviated as the PSO belongs to a kind of evolutionary algorithm, genetic algorithm, it is from a random solution by iteration to find the optimal solution, it is also to evaluatethequality of the solution through the fitness, but it is much simpler than the rules of thegenetic algorithm, it does not "cross" and "mutation" operation of genetic algorithms, which follow the optimal value of the current search to find the global optimum. This algorithm is its easy, high accuracy, fast convergence has attracted the attention of academia, and demonstrated its superiority in solving practical problems.In this paper, the characteristics of the particle swarm algorithm, application, elements of the formal definition introduced and summarized. One of the biggest advantages of the PSO algorithm do not need to adjust too many parameters, but a few parameters in the algorithm has a direct impact on the performance and convergence of the algorithm. At present, the theoretical study of the PSO algorithmis still in the initial stage, the algorithm parameter settings in a large extent also depends on the experience. PSO parameters include:particle number, particle length, particle range, the particle maximum rate, the inertia weight, learning factors. In this paper, the main parameters of the particle swarm algorithm and gave weight and focus on the learning factor and inertia weight in-depth analysis, given the derivation of specific mathematical formulas, given certain improvements to the learning factor parameters, on the inertia weight analyzed, and the latest results of theoretical studies. After the combination of the instance of the maximum rate of the particle and the inertia weight, re-combined in-depth analysis of the particle swarm algorithm, in order to test different maximum speed V_max of restrictions, the inertia weight factor affecting our choice of schaffer’s the f6standard test function. The final analysis concluded that the values of the inertia weight factor based on the iterative process changes conducive to the improvement of the algorithm.