| Abstract [eng] |
Agents controlled by an artificial intelligence have been used in games for a long time. Agent behaviour can be improved using optimization algorithms (such as genetic algorithm). To achieve a goal, an agent needs to understand his environment and actions made by other agents. The system implements the following methods for artificial intelligence: genetic algorithm, artificial neural network, influence maps. These methods are used to model actions made by an agent, and to coordinate those actions with other agents. They are applied specifically to solve agent related problems in a „Warlike“ turn-based game. The goal of this study is to compare the artificial intelligence methods used in a turn-based game. The results achieved in this study can be used to find balancing errors in similar turn-based games and to develop turn-based game artificial intelligence. |
