Abstract:
In recent years using real-time strategy (RTS) games as a test bed for Artificial Intelligence techniques has become increasingly popular. RTS games allow an environment that is complex enough to present the AI with a sufficiently challenging situation. These games also provide an easy way to evaluate the effectiveness of the technique. In this paper, I will describe how I built on the work of Aha, et al., who in [2] used a custom Case-Based Reasoner to govern an AI in Wargus, an open source RTS game, that learns to play better as it plays more games. I replicated their experiment with a different Case-Based Reasoner. I created and worked with BARC, a Case-Based Reasoner written in C++ for this specific application. BARC is based on the description of CAT in [2], but shares no source code.