Co-ordination and Co-operation in Agent Systems: Social Laws and Argumentation

The social laws paradigm represents an important approach to the co-ordination of behaviour in multi-agent systems. In this paper we examine the relationship between social laws and rational behaviour, by which we mean behaviour that can be justified by a defensible argument. We describe how social laws have previously been defined and used within the context of Action-Based Alternating Transition Systems (AATS). We then show how an account of argumentation for practical reasoning in agent systems, also based on AATS, can be used to determine what is rational for the agents to do in the absence and presence of such laws. The reasoning involved is both of a practical and epistemic nature: agents need to make decisions about what to do based upon the assumptions that they make about the states they find themselves in, and crucially, they also need to reason about what the other agents in the scenario will do. What is rational for the agents to do has implications for the need for social laws, the ways in which social laws can help the situation, the form the social laws should take, and the likelihood of compliance with the social laws. This paper demonstrates how we can think about social laws and rational behaviour in a single framework, so as to identify these implications in particular scenarios, and so frame social laws accordingly.