基于三角形模糊数的非线性T-S模糊系统的峰值点和分量半径优化

单值模糊器是将高维空间中一个实值点映射成该空间上的一个单值模糊集,在构造非线性T-S模糊系统时不仅可克服输入变量的噪声问题,而且能减少模糊推理机设计中的计算量.首先,基于分片线性函数和单值模糊器给出了非线性T-S模糊系统模型;并依据广义三角形的重心坐标公式,对等距剖分论域中的峰值点和分量半径等参数进行了优化;最后,通过模拟实例对系统进行了验证,得到优化后的非线性T-S模糊系统确实有更好的逼近效果.

Optimizations of peak points and branch radius of nonlinear T-S fuzzy system based on triangular fuzzy numbers

Single value fuzzifier is a mapping from a real value point to higher dimensional triangle fuzzy number in n-European space. It not only can overcome the noise of the input variables in constructing nonlinear T-S fuzzy system, but also can simplify the complicated calculation in the design of fuzzy inference engine. Firstly, a nonlinear T-S fuzzy system model is established based on the piecewise linear function and the single value fuzzifier. Secondly, the peak points and the branch radius in the equidistant subdivision universe are optimized by adopting the formula of barycenter of the generalized triangle. Finally, we verify that the optimized nonlinear T-S fuzzy system has good approximation effect by selecting the sample points.