摘要
提出了一种新的将隶属度函数和规则库统一编码的改进遗传算法进行TS模糊模型整体优化设计的方法。利用FCM算法和最小二乘法辨识初始的模糊模型;利用改进遗传算法整体优化模糊模型,克服了以往将模型结构和参数分开优化的缺陷。为了提高模型的解释性,提出了将基于相似性的模糊集合和模糊规则的简化方法用于对模型的约简,并利用该方法对Mackey-Glass混沌序列建模。仿真结果验证了该方法的有效性。
A new improved genetic algorithm which combined membership function and rules sets in a chromosome is proposed. The initial fuzzy system is defined using fuzzy clustering algorithm and the least-squares. Improved genetic algorithm is used to optimize the whole system,conquer the usual method＇ s limination which the model structure and parameter is optimized each. Similar fuzzy sets merging and fuzzy rules mergering are adopted to reduce the fuzzy model and enhance its interpretablity. The proposed approach is applied to the Mackey-Glass system, and the results show its validity.
出处
《传感器与微系统》
CSCD
北大核心
2008年第8期100-102,120,共4页
Transducer and Microsystem Technologies
基金
江苏省国际合作计划资助项目(BZ2005035)
关键词
TS模糊模型
模糊聚类
改进的遗传算法
解释性
精确性
TS fuzzy model
fuzzy clustering
improved genetic algorithm
interpretability
precision
