首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种新的元胞网络模型
引用本文:李学伟,孙有发,吴今培.一种新的元胞网络模型[J].五邑大学学报(自然科学版),2011(4):14-21.
作者姓名:李学伟  孙有发  吴今培
作者单位:1. 大连交通大学,辽宁大连,116028
2. 广东工业大学管理学院,广东广州,510520
3. 五邑大学智能技术与系统研究所,广东江门,529020
基金项目:国家自然科学基金资助项目
摘    要:用元胞替换传统人工神经网络中的神经元,以局部连接取代相邻层级元胞之间的全连接,用规则演化算法替代BP算法,建立元胞网络模型.设计了元胞网络的训练过程:内嵌的遗传算法用于寻优各元胞状态离散化边界值以及元胞网络拓扑结构.以一个红酒分类基准数据为例,检验了元胞网络的学习性能和分类性能,获得了较为满意的结果.

关 键 词:元胞自动机  神经网络  遗传算法  规则演化  元胞网络

A New Kind of Cellular Network
LI Xue-wei,SUN You-fa,WU Jin-pei.A New Kind of Cellular Network[J].Journal of Wuyi University(Natural Science Edition),2011(4):14-21.
Authors:LI Xue-wei  SUN You-fa  WU Jin-pei
Institution:1.Dalian Jiaotong University,Dalian 116028,China;2.School of Management,Guangdong University of Technology,Guangzhou 510520,China;3.Institute of Intelligence Technology and Systems, Wuyi University,Jiangmen 529020,China)
Abstract:A novel concept of cellular network is proposed. The cellular network is constructed based on the traditional artificial neural network by replacing the neurons with ceils, replacing the full connections among cells with partial connections, and replacing the back-propagation algorithm with rule-evolution algorithm. The training process of cellular network is also designed. The embedded genetic algorithm is to search for the optimal parameters and structures of the network. Finally, the cellular network is applied to a benchmark data set (classification of Italy's red wine), and is validated to be of good performances of learning and predicting.
Keywords:cellular automata  neural network  genetic algorithm  rule evolution  cellular network
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号