动量交叉粒子群算法

针对传统粒子群算法收敛速度慢、全局搜索能力差的缺点,提出了一种新的最优交叉动量粒子群算法。该算法通过在粒子群算法中引入一种新的二进制交叉策略来加快粒子群算法的收敛速度,通过设置新的惯性权重来改善新算法的全局搜索和局部搜索能力,并且在粒子搜索路径中引入变系数低通滤波器构成的动量算法来平滑粒子搜索路径。几个典型的测试函数仿真表明,新算法在收敛速度和搜索精度上均得到了明显改善。

Momentum particle swarm optimization with optimal crossover

Aiming at the PSO's shortcoming about slow convergence rate and badly global searching ability, this paper presented a new particle swarm optimization with optimal crossoverOCPSO. By introducing a new simulated binary-crossover strategy SBX and a new strategy of inertia weight setting, it improved the ability of global and local searching. Furthermore, it utilized variable coefficient low-pass filters to update particles' positions of OCPSO, called momentum algorithm, which could enhance the speed and accuracy of convergence. Experimental results on several classical functions indicate that the new algorithm can greatly improve the searching speed and accuracy.