二阶微粒群算法

为了提高标准微粒群算法的全局收敛性,提出了一种新的微粒群算法——二阶微粒群算法.首先,介绍了二阶微粒群算法的引入,分析了其收敛性,并且研究了其参数的选择范围.其次,在分析二阶微粒群算法的进化方程的基础上,引出了具有随机惯性权重的标准微粒群算法.再次,在二阶微粒群算法中加入振荡因子来调整微粒的速度变化率,更好地使二阶微粒群算法收敛于全局最优.最后,利用这几种改进方法对典型测试函数进行仿真,实验结果表明,这些方法能够有效克服早熟问题,在全局收敛性和收敛速度方面均优于标准微粒群算法.

A Two-Order Particle Swarm Optimization Model

Particle swarm optimization (PSO) is an evolutionary computation technique developed by Kennedy and Eberhart in 1995. The underlying motivation for the development of PSO algorithm is social behavior of animals such as bird flocking, fish schooling, and swarm theory. Now PSO has been proved to be very effective for some problems. However, like other stochastic algorithms, PSO also suffers from the premature convergence problem, especially in the large scale and complex problems. In order to improve the global convergent ability of the standard particle swarm optimization (SPSO), a new version of particle swarm optimization named by a two-order PSO model is developed. Firstly, a two-order PSO model is introduced, its convergence analysis is given, and at the same time its parameter choices are studied. Secondly, a PSO model with the stochastic inertia weight is educed from the evolutionary equations of the two-order PSO. Thirdly, the two-order oscillating PSO with an oscillating factor is provided to adjust the influence of the acceleration on the velocity, which can guarantee the two-order PSO to converge to the global optimization validly. Finally, the above-proposed models are used to some benchmark optimizations. The experimental results show the proposed models can overcome the premature problem validly, and outperform the standard PSO in the global search ability and convergent speed.