基于模糊竞争学习和UD分解的模糊建模

Fuzzy modeling based on fuzzy competitive learning and matrix UD decomposition

为解决复杂系统模糊建模问题,研究了利用模糊竞争学习实现非线性系统的模糊建模方法.首先,利用模糊竞争学习方法划分输入变量的模糊输入空间,确定模糊模型的规则数、规则,实现模糊模型的结构优化.另外,为了克服递推最小二乘出现的误差积累、传递现象,采用基于矩阵UD分解的递推最小二乘方法确定模糊模型的结论参数,从而实现模糊模型的结构和参数优化.采用该方法对M ackey-G lass混沌时间序列进行建模研究,结果表明可以在线或者离线对M ackey-G lass混沌时间序列进行准确预测,效果较好.

For the fuzzy modeling problem of complex system, the fuzzy modeling of nonlinear systems based on fuzzy competitive learning is proposed. First of all, the fuzzy competitive learning is utilized to partition the input space of input variables, and to confirm the number of rules and rules, and then optimize the structure of fuzzy model. In addition, the recursive least square based on matrix UD decomposition is used to confirm the conclusion parameters of fuzzy model for the sake of accumulating and transferring of the errors of recursive least square. The structure and parameters of fuzzy model are optimized on the basis of the presented algorithm. To illustrate the performance of the proposed method, simulations on the chaotic Mackey-Glass time series prediction are performed. Combining either off-line or on-line learning with the proposed method, the chaotic Mackey-Glass time series are accurately predicted, and the good effectiveness is demonstrated.