摘要
针对基本粒子群算法(PSO)易陷入局部极值,后期迭代效率不高的缺点,提出了一种简化的位置随机扰动粒子群算法(SPSDPSO)。新算法通过取消粒子的速度项改善了算法的收敛性能,同时,在满足种群多样性的评价指标阈值的条件下引入粒子位置的随机扰动,提高了算法的全局收敛性,减少了算法陷入局部极小值的可能。仿真表明,新算法具有较强的全局搜索能力,而且能有效地避免常规算法的早熟收敛问题,显著的提高了优化性能。
A simple position stochastic disturbance particle swarm optimization algorithm (SPSDPSO) was presented aimed at the basic PSO algorithm's disadvantages such as getting local optimization easily and low iterative efficiency. The new algorithm improved con- vergence capability by canceling the particle's velocity, at the same time, imported position stochastic disturbance to improve the algorithm's convergence capability and reduce likelihood on getting into local optimization. The simulation result showed that the new algorithm could greatlv improve the global convergence ability and was an effective improved particle swarm optimization.
出处
《微计算机信息》
2009年第18期218-219,217,共3页
Control & Automation