Law-governed multi-agent systems: from anarchy to order

I will argue that a multi-agent system should not be viewed as just the totality of its members, but that it should include the set of laws that governs the interaction between these members. Such laws need to be made explicit, as part of the design of the MAS, and they need to be enforced on its implementation.I will discuss the architecture and the scalable implementation of such law-governed MAS, using a coordination and control mechanism called Law-Governed Interaction (LGI); and will consider various implication of this architecture, to issues such as reasoning about a MAS, and making it dependable.