基于多样性变异的量子行为粒子群优化算法

Quantum-behaved particle swarm optimization with diversity-guided mutation

为了克服量子行为的粒子群优化(QPSO)算法存在早熟收敛的缺点,提出了一种改进的QPSO算法,在QPSO算法中加入多样性变异算法、设置多样性函数,当多样性较少时,执行变异操作;扩大了种群搜索过程中的搜索范围,避免了种群多样性不断下降。典型标准函数优化的仿真结果表明,该算法具有较强的全局搜索能力。

To overcome the premature convergence of quantum-behaved particle swarm optimization（QPSO） algorithm,this paper proposed QPSO with diversity-guided mutation（QPSO-DGM） to improve the performance of QPSO.In the proposed QPSO-DGM algorithm,set diversity function.When the value of diversity was less during the search,operated the mutation.QPSO-DGM made the particles＇ search scope expanded and avoided the declination of population diversity.The experiment results on benchmark functions show that both QPSO-DGM have stronger global search ability than QPSO and standard PSO.