λ-模糊测度辨识方法研究

A Study on Identification Methods of λ-Fuzzy Measures

研究了当专家给出的模糊测度具有不确定性时λ-模糊测度的辨识问题 ,不确定数据分别用区间和模糊数来表达 ,针对不同情况提出了 4种误差函数 ,误差函数的最小值对应于最优的辨识结果 ,而且当不存在不确定性时 ,所得结果与目前已有方法相同。数值实验结果表明 。

Fuzzy measures given by human experts usually contain uncertainties, and do not satisfy required mathematical properties. The problem of identifying lambda fuzzy measures is studied when the uncertainties are expressed by intervals and fuzzy numbers. Four error functions are proposed. The minima of error functions correspond to optimal identified results. When there is not any uncertainty, obtained results are equal to those from existing methods. Numerical experimental results indicate that the proposed identification algorithms are efficient.