核矩阵协同进化的震荡搜索粒子群优化算法

Shock search particle swarm optimization algorithm based on kernel matrix synergistic evolution

针对粒子群算法搜索后期易陷入局部极值的缺点,提出一种基于核矩阵协同进化的震荡搜索粒子群优化(kenel matrix synergistic evolution shock search particle swarm optimization,KMSESPSO)算法,该算法对粒子进行局部与全局结合的震荡搜索,且当整个粒子种群陷入停滞状态时,利用核矩阵对特定粒子组进行协同进化以扩大种群的多样性。实验结果表明,KMSESPSO算法有效提高了粒子的全局搜索能力,既避免粒子种群易早熟收敛,又较好地提高寻优精度、加快收敛速度,且有一定的鲁棒性。

Due to the shortcoming of particle swarm optimization（ PSO） algorithm that it is often trapping in local optimum at the late stage,a kind of shock search PSO algorithm based on kernal matrix synergistic evolution（ KMSESPSO） is proposed. The proposed algorithm does a combination of local and global shocks search and when the whole particle swarm is stagnant a specific particle group would have a synergistic evolution to enrich the diversity of population by using kernel matrix. The experiment results show that the proposed algorithm strengthens the global search capability of particles effectively and can not only get free from premature but also raise the optimal accuracy in faster convergence speed and have certain robustness.