Applying fuzzy set theory and case-based reasoning approach for managing strategical and tactical reasoning in starcraft.
Export citation
Abstract
This thesis aims at describing, analyzing, implementing and discussing fuzzy set theory and Case-Based Reasoning (CBR) for strategical and tactical management in the real-time strategy game of StarCraft. In order to play a complete match of StarCraft this thesis divides the problem in four categories: resource management, strategical management, tactical management and micro organization. Case-based reasoning is a problem solving AI approach that uses past experience to deal with actual problems. A new problem is solved by finding a similar past case, and reusing it in the new problem situation. Fuzzy set theory is used in case representation to provide a characterization of imprecise and uncertain information. In this thesis, the combination of fuzzy sets and case-based reasoning is called Fuzzy Case-Based Reasoning (FCBR). CBR was applied to reason about strategies while FCBR was applied to deal with tactical reasoning. The resulting system was victorious in 60% of the games, it was defeated in 25% of the games, getting ties in 15% of the matches. The results revealed that our system can successfully reason about strategies and tactics, defeating the built-in AI of StarCraft. The principal conclusion was that FCBR can reason with abstract information and a large space of actions. Moreover, the resulting system shows its potential to incorporates human knowledge and can effectively adapt to varying conditions of the map.