一种混沌混合粒子群优化RBF神经网络算法  

An Algorithm of Chaotic Hybrid Particle Swarm Optimization Based on RBF Neural Network

作　　者：刘洁[1] 李目[2] 周少武[2] 

机构地区：[1]湖南工程学院设计艺术学院,湖南湘潭411104 [2]湖南科技大学信息与电气工程学院,湖南湘潭411201

出　　处：《计算机技术与发展》2013年第8期181-184,共4页Computer Technology and Development

基　　金：国家自然科学基金资助项目(60673119);国家863/CIMS项目(2006AA04Z152);湖南省科技计划项目(2006GK3071);湖南省教育科研项目(10C0672)

摘　　要：为了更精确地检测出混沌背景下的微弱目标信号,提高预测效果,文中提出了一种混沌混合粒子群优化RBF神经网络(CHPSO-RBFNN)算法。本算法主要采用了基于群体自适应变异和个体退火操作的混沌粒子群优化RBF神经网络,利用群体自适应变异以及个体退火操作优化混沌粒子群,有效地提高了粒子群算法的全局收敛性,优化了RBF神经网络的结构和参数。把该算法用于预测混沌时间序列、检测混沌背景下微弱目标信号,实验结果表明本算法有良好的非线性预测能力,可以有效地检测出混沌背景下的微弱目标信号。In order to detect the weak target signal accurately in the chaos background, and improve forecast result, a novel algorithm based on RBF Neural Network （RBFNN） with Chaotic Hybrid Particle Swarm Optimization （CHPSO） is presented. In this algorithrn, the RBF neural network is optimized by chaotic particle swarm optimization with adaptive population mutation and individual annealing operation. In order to improve the global convergence ability of PSO, the colony adaptive mutation and individual annealing operation are used to adjust and optimize PSO. Then the parameters and structures of RBFNN are optimized. This novel algorithm is applied to predict chaotic time sequence and detect weak target signal in the chaos background. Simulation results show that the algorithm has preferable nonlinear prediction ability and can detect weak target signal effectively.

关 键 词：混沌 自适应变异 粒子群 模拟退火 RBF神经网络 目标检测 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]