摘要
为提高粒子群算法的收敛性,基于动力系统的稳定性理论分析了带有惯性权重的粒子群算法模型的收敛性,提出了在算法模型收敛条件下惯性权重w和加速系数c的参数约束关系。使用4个测试函数对具有所提参数约束关系的惯性权重粒子群算法模型和典型参数取值惯性权重粒子群算法模型进行了对比仿真研究,实验结果表明,具有提出的参数约束关系的惯性权重粒子群算法模型在收敛性方面具有显著优越性。
In order to improve the convergence of particle swarm optimization （PSO）, convergence performance of the PSO with inertia weight （IPSO） is analyzed based on stability theorem of dynamic system. And the constraint relationship between the acceleration coefficient c and the inertia weight w is proposed to ensure the convergence of the IPSO model. The IPSO satisfying the performed w-c constraint condition is tested with four well-known benchmark functions compared with IPSO model with typical values ofw and c. The experimental results show that IPSO model satisfying the proposed w-c constraint condition has better convergence performance.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第18期4068-4071,共4页
Computer Engineering and Design
基金
江苏高校自然科学基金项目(08KJD510011)
关键词
粒子群算法
动力系统稳定性理论
惯性权重
加速系数
收敛性
PSO
stability theorem of dynamic system
inertia weight
acceleration coefficient
convergence
