The workshop will have four main themes:


1. Technical hnical

What techniques are suitable for agents that are incorporated in educational contexts, games and simulations. How to balance intelligence and efficiency? How to couple the agents to the game/simulation and manage this coupling’s information flow? How to deal with the inherent real time nature of the game engine environment? How to couple long and short time interactions?


2. Conceptual

What information is available for the agents' use, either through the educational context, or from the system, through for example, the game or simulation engine? How can reaction to events be balanced with with goal directed behaviour? How are ontological differences between information used by agents and information from the domain handled? How do we choose the actions of an agent? Too high level gives little control; too low level makes the agent inefficient.


3. Design

How to design interactive systems containing intelligent agents. How do we determine what agents should do and should not do, such that local autonomy and story line are well balanced. How to design the agents themselves that are embedded in other (possibly diverse) systems (including the behaviour authoring tools and methodologies).


4. Education

It is also important that we introduce both the design and construction of these collaborative autonomous systems into the computer science curriculum and develop ways of encouraging their effective utilisation across the curriculum. Contributions to the workshop will be welcomed that provide a mixture of relevant theoretical and practical understanding of both the teaching and use of multi-agent systems in educational and entertainment research, together with practical examples of the use of such systems in real application scenarios. These will be written for students, teachers, producers, directors and other professionals who want to improve their understanding of the opportunities offered by the use of multi-agent systems in teaching and entertainment scenarios of all types.