摘要
针对神经模糊系统中获取控制规则方法———模糊聚类法的不足提出了改进措施. 改进的方法在聚类半径和聚类中心的选取上更趋于合理. 通过仿真比较了原算法和改进算法对样本的拟合,结果表明了改进算法的有效性,依据改进方法获得的控制规则设计的神经模糊系统,具有较强的自适应能力.
In view of the shortcomings in obtaining the fuzzy rules on fuzzy neural network system, we suggest some improvement on the method. The improved method is more reasonable than ever before so far as the choice of clustering radius and focus is concerned. The simulation result proves the validity of the improved method compared with the old one. The neural fuzzy system designed with the improved control rules is more self-adaptive.
出处
《大庆石油学院学报》
CAS
北大核心
2005年第1期85-88,共4页
Journal of Daqing Petroleum Institute
关键词
神经网络
模糊控制
模糊规则
聚类
优化
neural network
fuzzy control
fuzzy rules
clustering
optimization
