An attitude based teamwork model for a hostile dynamic multiagent world

In this thesis, we propose a teamwork model based on the concept of a mental attribute called attitude. Our team model presents a novel way of solving problems and conflicts in our domain. We argue that this collective attitude is further decomposed into the individual attitudes of the agents towards various team attributes. We also present a team problem solving methodology, which has the notion attitude and team cycle as its core to allow robust and coherent team behaviour. The application and implementation of this model and methodology to a virtual fire-fighting domain has revealed a promising prospect in developing team agents. We also suggest various team and fire attitudes for the agents, which can be very useful for a hostile and dynamic world. We then evaluate the team problem solving behaviours of the agents in a simulated fire world using teams with and without different types of attitudes.