|Master:||project within::Cognitive Science|
|Student name:||student name::Daniël Höhle|
In this thesis I will suggest a revised form of reasoning for agents in virtual environments, eg. games. Nowadays, agents usually reason rationally, without taking their emotions into account. Here I will present an upgraded form of the BDI model which also incorporates capabilities and emotions. The purpose of my research is to give agents more human-like behaviour, where the agent's action is dependent on the 'mood' and the emotions the agent is experiencing with respect to the environment and other agents. The project will be addressed in two steps. First, a conceptual model will be designed and tested in the LeadsTo simulation environment. Second, this model will be implemented within virtual players in the RoboCup simulation environment. For the RoboCup simulation I will use existing team codes, used in previous competition settings, and adjust the decision structure in a way that is consistent with the improved BDI model. After proving the relevance of decision making with emotions and trust, this model could be tested and implemented in various agent settings, from First-Player-Shooters to team-based gaming.