简化模糊推论模型与多层构造模糊推论模型的对比研究

Comparison with the fuzzy reasoning model of single layer and multi-layer structure

能够直接输出非模糊值的简化模糊推论已被广泛应用于控制系统 .随着条件部成员函数个数的增加 ,这种模糊推论模型能够获得较高的输出精度 .但是 ,对于采用从现实环境下获取的示教数据进行模型学习的情况 ,简化模糊推论模型有时难以获得很高的辨识精度 .本文采用现实环境下的两种典型的示教数据 ,对单层构造的模糊推论模型和多层构造的模糊推论模型进行辨识 .实验结果表明 ,多层构造的模糊推论模型比单层构造的模糊推论模型具有更高的辨识精度和对示教数据的适应能力 .图 7,表 4,参

The simplified fuzzy reasoning model with the advantage that the model can output real number directly are frequently applied to the field of the control system. And the model can generate with higher accuracy by increasing the number of fuzzy partitions in antecedent. However when the teaching data is obtained from actual environment, it is difficult to generate the fuzzy reasoning model with high identification accuracy.This paper identifies the simplified fuzzy reasoning model and the multi-layer structure fuzzy reasoning model, using 2 kinds of typical teaching data abstracting from actual environment. Through identification experiments, it shows that the multi-layer structure fuzzy reasoning model has higher identification accuracy, and stockier adaptability to the teaching data than single layer structure fuzzy reasoning model.7figs.,4tabs.,16refs.