Fuzzy-Based Decision Strategy in Real-Time Strategic Games

The aim of this article is to describe our own gaming artificial intelligence for OpenTTD, which is a real-time building strategy game. A multi-agent system with fuzzy decision-making was used for the proposal itself. The multi-agent system was chosen because real-time strategy games achieve great complexity and require decomposition of the problem into individual problems, which are then solved by individual cooperating agents. The system becomes then more stable and easily expandable. The fuzzy approach makes the decision-making process of strategies easier thanks to the use of uncertainty. In the conclusion, own experimental results were compared with other approaches.