模糊類神經網路之函數近似使用基因演算法

Abstract

本文中我們利用類神經網路的學習能力，建立一個擁有自動調整權重值的模糊推論系統。而基因演算法是一種族群搜尋策略的最佳化方法，所以我們使用基因演算法去搜尋模糊類神經網路中最佳的權重值w/sub i/來近似函數。然而基因演算法中最大的缺點在於搜尋到最佳值的速度太慢，因此在本文中提出新的交配運算方法並且探討其對搜尋結果之影響。

In this paper, we use the learning ability of neural networks to builda fuzzy inference system, in which weighting factor values areautomatically adjusted by genetic algorithms. Because of thesuperiority of genetic algorithms in directed random search for globaloptimization, they are used to obtain a set of optimal fitnessweighting values for the fuzzy neural networks to approximatefunctions to desired accuracy. To address the problem oftime-consuming evolutionary process in searching for an optimum value,we propose a novel crossover methodology for the genetic algorithm sothat the system performance can be improved. Also, the effect of theproposed crossover methodology on searching results is alsoinvestigated in the paper.