摘要
为克服混沌粒子群优化(CPSO)算法由于采用随机数作为算法参数而不能保证种群多样性和优化遍历性的缺陷,通过将混沌变量嵌入到常规粒子群优化算法(PSO)中,使PSO算法中的惯性权值和随机数用混沌随机序列来替代,提出了一种新的混沌粒子群嵌入优化算法(CEPSO),以充分利用混沌运动的随机性、遍历性克服粒子群优化算法容易陷入局部最优的缺点。通过复杂多维函数的寻优测试,验证了本算法的有效性,并将仿真结果与混沌粒子群优化算法进行比较,证明了CEPSO算法更具有较强的全局搜索能力和收敛速度。
The chaos particle swarm optimization ulation multiplicity and the optimized ergodicity, (CPSO) algorithm cannot guarantee the pop- because it uses the random number as the al- gorithm parameter. This paper proposes a new chaos embedded particle swarm optimization (CEPSO) algorithm and uses chaotic maps to substitute the inertia weight and random num- bers of the classical PSO algorithm. The stochastic property and the ergodicity of chaotic search are used to overcome the shortcoming of PSO algorithm trapped in local optima. Experi- ments on complex functions with more dimension demonstrate that the CEPSO algorithm out- performs original CPSO on the global searching ability and at the convergence speed.
出处
《数据采集与处理》
CSCD
北大核心
2010年第1期102-106,共5页
Journal of Data Acquisition and Processing
基金
上海市教委科技发展基金(050Z02)资助项目
上海应用技术学院科技基金(KJ2009-12)资助项目
关键词
嵌入优化算法
粒子群
混沌
全局最优
embedded optimization algorithm
particle swarm
chaos
global optimization