摘要
本文在分析判决反馈RBF网络信道均衡器(DFRBFNE)的结构特点基础上,定义一个布尔向量L作为网络的结构参数,与原来DFRBFNE的隐节点参数集一起构成了新的DFRBFNE隐节点参数集{c,σ,L},并给出了一个新的DFRBFNE网络输入输出关系表达式;采用一种混合协同微粒群算法同时对DFRBFNE网络拓扑结构和隐层节点参数进行优化设计,并将输出线性参数集分离后采用最小二乘法进行优化设计,简化了优化空间,加速了算法的收敛速度。
According to the characteristics of the RBFN equalizers with decision feedback, a boolean variable L is defined as the network structural parameter and is combined with the original reasoning parameter set to form a new parameter set {c, a, L}, and a new expression related input and output of ANFIS is obtained. Then a hybrid cooperative particle swarm optimization algorithm is proposed to optimize the parameter set {c, a, L}, and the output linear parameter set {w} is separated to optimize using LMS. In the proposed method, the dimension of the operating space is reduced and the algorithm convergence speed is increased by the hybrid PSO.
出处
《计算机工程与科学》
CSCD
2008年第6期79-82,共4页
Computer Engineering & Science
基金
国家自然科学基金资助项目(60673119)
湖南省教育厅优秀青年科研资助项目(07B019)
湖南省教育厅科研资助项目(06C337)
关键词
信道均衡器
RBF网络结构
微粒群算法
混合协同
channel equalizer
RBF network structure
particle swarm optimization
hybrid cooperative PSO
