作　　者： ; ;

机构地区： 重庆教育学院计算机与现代教育技术系

出　　处： 《重庆教育学院学报》 2007年第3期42-45,共4页

摘　　要： 本文采用集中性和多样性策略对禁忌搜索进行改进,提出了一种基于模糊神经网络的混合禁忌搜索优化算法(FNN-based Hybrid Tabu Search Algorithm,FNN-HTS),用于同时优化模糊神经网络的结构和参数以提取出一组尽量精练的模糊规则。在FNN-HTS中,禁忌搜索用于同时优化网络结构和隶属函数参数,结合最小二乘法快速求解规则后件的线性参数。非线性函数逼近的实验结果表明所提出的方法能获得一组更精练的规则和更小的误差。 In this paper,a hybrid Tabu Search algorithm based on fuzzy neural network （FNN-HTS） was proposed to generate an appropriate fuzzy rule set automatically through structure and parameters optimization of fuzzy neural network ,where an intensification and diversification strategy was used to improve performance of primitive Tabu Search. In FNN-HTS,Tabu Search was used to optimize the network structure and membership function simultaneously,after which,Least Squares was used for the consequent parameters of the fuzzy rules. A simulation for a nonlinear function approximation was presented and the experimental results showed that the proposed algorithm can generate more refined rules with a lower average percentage error.

关 键 词： 禁忌搜索 模糊神经网络 函数逼近

领　　域： [自动化与计算机技术] [自动化与计算机技术]