Teamwork via team plans in intelligent autonomous agent systems

Abstract

Intelligent agent systems are developing into an important paradigm in Artificial Intelligence and in general computer science. An agent can be seen as an autonomous system that perceives and acts on its environment, performing tasks and pursuing goals independently of user control. Recently, there has been much research interest directed toward the design of multi-agent systems that act collaboratively on tasks that are too complex for any single agent to perform on its own. We discuss the design and implementation of teamwork amongst autonomous agent systems. In particular, we describe an approach, using team plans, that addresses a number of important issues, such as: balancing autonomous behaviour with commitment to team behaviour; team formation; distribution of tasks amongst team members; coordination and synchronisation of actions.