一种融合差分进化的量子粒子群优化算法

Quantum Behaved Particle Swarm Optimization Algorithm Merging Differential Evolution

由于量子粒子群优化算法在迭代后期易出现粒子多样性差,收敛到局部最优解等缺点。为解决上述问题,提出一种融合差分进化的量子粒子群优化算法。上述算法在量子粒子群算法的上,首先利用差分思想对粒子的速度提出一种改进策略,再对陷入局部最优的粒子进行交叉选择操作,从而较好的保持种群中粒子的多样性,避免粒子后期陷入局部最优。通过对3个测试函数进行的仿真,结果表明融合了差分进化算法的量子粒子群算法具有收敛速度快、收敛能力强等特点,解决了算法局部最优问题。

To overcome the problem that quantum particle swarm optimization algorithm may lost the particle diversity and converge to local optimal solution in the later phase of the iteration,a quantum particle swarm optimization algorithm merged with differential evolution was proposed. Firstly,an improved strategy was proposed to improve the speed of the particles based on the idea of differential algorithm. And then,crossover and selection operations were involved into quantum particle swarm algorithm. Therefore,it can better keep the diversity of the particles in the population,and avoid the local optimum in the later phase of the iteration. Through the simulation experiments of 3test functions,the results show that the quantum particle swarm algorithm by integrating differential evolution algorithm has the characteristics of fast convergence speed,strong convergence ability and so on.