人工生命中以模糊法則為基礎的行為學習之研究

Abstract

在人工生命的領域中，行為的選擇與表現是最重要的一環。在本篇論文中，我們利用模糊理論來加速與簡化行為的選擇，讓行為的選擇能夠更一般化，並且加入行為的學習能力。我們同時也提出了一個改良後的系統架構。應用我們的架構，將可以更容易的建構出豐富的人工生命。

Behavior-based modeling is the most popular technique to model autonomous creatures in artificial life. In this thesis, we propose a behavior-based architecture integrating fuzzy theory to facilitate the learning of weights of complex behavior networks and make the learning of weights more general. With this architecture, our autonomous creatures will learn and select appropriated behaviors on their own in artificial life.