您的位置:首页  > 论文页面

混沌优化算法在多目标数据关联中的应用

发表时间:2009-09-30  浏览量:1129  下载量:418
全部作者: 李冰玉,郝芸
作者单位: 天津理工大学中环信息学院
摘 要: 针对Hopfield神经网络中输出函数的神经元增益1/u0对网络收敛动力学特性的影响,以及混沌神经网络中控制自反馈连接权衰减速度的参数β对网络混沌动态特性的影响,提出在混沌神经网络模型输出函数中引入时变的神经元增益系数1/[λ+εi(t)], 以修正自反馈连接权的表达式,丰富神经网络的动态特性。仿真实验结果证明:优化的混沌神经网络在获得较好的混沌搜索能力的同时,也加快了原有算法的稳定收敛速度。
关 键 词: 数据处理;Hopfield神经网络;混沌神经网络;神经元增益;自反馈连接权
Title: Application of chaos optimization algorithm in multi-target data association
Author: LI Bingyu, HAO Yun
Organization: Zhonghuan Information College, Tianjin University of Technology
Abstract: Aiming at the influences of neuron gain 1/u0 of the output function in Hopfield neural network on the convergent dynamic characteristics of the network, and the influences of parameter β used for controlling decay rate of self-feedback connecting weight on the chaos dynamic characteristics of the network in the chaos network, the paper proposes to introduce a time-variant neuron gain factor 1/[λ+εi(t)] into the output function of the chaos neural network model for correcting expression of the self-feedback connecting weight, so as to enrich the dynamic characteristics of the neural network. The simulation results show that the optimization chaos neural network gets better chaos searching ability and speeds up the stability convergence rate of the original algorithm at the same time.
Key words: data processing; Hopfield neural network; chaos neural network; neuron gain; self-feedback connecting weight
发表期数: 2009年9月第18期
引用格式: 李冰玉,郝芸. 混沌优化算法在多目标数据关联中的应用[J]. 中国科技论文在线精品论文,2009,2(18):1931-1936.
 
0 评论数 0
暂无评论
友情链接