摘要
适应性粒子群寻优算法I(APSO-Ⅰ)是在有序的决策中始终引入随机的、不可预测的决定.为解决APSO-I算法收敛深度不够的问题,提出适应性粒子群寻优第II代算法(APSO-Ⅱ).APSO-Ⅱ算法是将有序(标准PSO粒子群寻优)和无序(自适应寻优)进行适当的分离,以发挥各自的优势.在自适应寻优阶段,通过在最优粒子邻域空间探寻更优化的解.一但新的优化解被发掘,便利用标准PSO快速寻优.典型复杂函数优化的仿真结果表明,APSO-Ⅱ在收敛速度和收敛深度上均优于DPSO(耗散型PSO),HPSO(自适应层次PSO),AEPSO(自适应逃逸PSO)和APSO-Ⅰ.
Adaptive particle swarm optimisation-Ⅰ (APSO-Ⅰ) simulates the complex behaviour of social swarm and overlaps the inscrutable decision on the rational behaviour. APSO-Ⅱ is proposed to overcome the poor convergence depth of APSO-Ⅰ. The APSO-Ⅱ algorithm divides the order action (the standard PSO) and the random exploration ( the adaptive optimisation) to show the advantage of two optimisation methods. In the stage of adaptive optimisation, the optimal solution is searched in the adjacent space of the best particle. Once the optimal solution is found, the standard PSO will be applied to rapidly explore. Experimental simulations show that APSO-Ⅱ algorithm is better than DPSO(Dissipation PSO), HPSO(Hierarchical PSO), AEPSO(Adaptive escape PSO), and APSO-Ⅰ algorithms in the capacity of convergence speed and convergence depth.
出处
《控制与决策》
EI
CSCD
北大核心
2009年第6期859-863,共5页
Control and Decision
基金
重庆市自然科学基金项目(CSTC2008BB6163)
关键词
粒子群算法
适应性
随机
收敛
Particle swarm optimisation(PSO)
Adaptive
Random
Convergence