基于减法聚类模糊神经网络高速公路混沌控制  

Chaos Control of Freeway Based on Fuzzy-neural Networks with Subtractive Clustering

作　　者：庞明宝[1] 贺国光[2] 任沙沙[1] 

机构地区：[1]河北工业大学土木工程学院,天津300401 [2]天津大学系统工程研究所,天津300072

出　　处：《武汉理工大学学报（交通科学与工程版）》2010年第5期1018-1021,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目(批准号:50478088;50808064);天津市科技支撑计划重点项目(批准号:08ZCKFSF01100)资助

摘　　要：研究了基于减法聚类的高速公路混沌系统模糊神经网络控制方法.提出通过数据挖掘技术建立交通流混沌控制器知识库的思想,设计了以密度、上游流量和最大李亚普诺夫指数作为输入,红灯时间作为输出的T-S模糊神经网络混沌控制器.采用减法聚类确定控制器结构提取模糊规则、控制器初始参数;应用模糊神经网络方法对控制器参数进行优化;结合遗传算法对聚类半径进行优化.仿真实验分析了该控制方法的控制效果,证明了该混沌控制方法的有效性.The chaos control problem of freeway was studied based on fuzzy-neural networks with subtractive clustering.The thought was proposed to establish the knowledge base of the chaos controller in freeway by using data mining technology.The fuzzy T-S ramp controller was designed.The traffic density,upstream traffic volume and maximal Lyapunov exponent are the input variables and the red timing is the output variable of the controller.Subtractive clustering is used to determine the controller structure including the extracting of fuzzy rules and generating initial parameters.The parameters of the fuzzy controller were optimized by using fuzzy-neural networks.The radius of the clustering centers was determined by genetic algorithm.The effects are discussed in the simulation experiment. The results indicate that the method is correct and feasible.

关 键 词：高速公路 混沌控制 T-S模糊神经网络 减法聚类 遗传算法 

分 类 号：U491[交通运输工程—交通运输规划与管理]