Abstract

this paper is to provide a framework in which one particular class of social activity can be formalised and ultimately analysed: namely that in which a group of autonomous agents (at least two) decides they wish to work together as a team to solve a common problem. A comprehensive theory describing this class of social interaction would need to cover at least the following aspects: when to initiate team activity, how to go about assembling the team, how to plan and distribute work within the team, how to behave once team activity has been initiated and how to complete team activity. The framework described herein defines the prerequisites for such action and also prescribes how agents should behave (both in their own problem solving and with respect to other group members) once the problem solving has been established. Typically in a community of autonomous agents, one of the primary motives for joint action is when no individual is capable of achieving a desired objective alone; only by combining and coordinating with others can the target be reached. Joint action is usually a reciprocal process in which participating agents augment their objectives and problem solving to comply with those of others - hence it is a fairly sophisticated form of cooperation. It requires greater knowledge, awareness and reflection by an agent both with respect to its own problem solving objectives and about their compatibility with the objectives of others, than simpler forms of social interaction (such as task and result sharing [19]). Joint action, by definition, requires an objective the group wishes to achieve - it is the glue which binds the team together. As a consequence of the autonomous nature of the agents, team members will only participate if they can derive some benefit from ...

Citations