Representing and enforcing interaction protocols in multi-agent systems: an approach based on conceptual graphs

Abstract

For a multi-agent system to work according to its specifications, the individual agents that compose it must abide by the interaction protocol that binds their actions. This interaction protocol delimits what the agents can do, so therefore, their ability to communicate, negotiate, and collaborate. By making some actions conditional to some other actions, it even enforces the control structure of the system. This paper presents a simple representation based on conceptual graphs used to describe an interaction protocol in a declarative way, and shows how subsumption between constraint graphs and asserted graphs can be used to enforce it. Our claim is that a declarative, logic-based, graphical representation of an interaction protocol will help the syst em designer to describe and validate it, to enforce it, and subsequently to update it, therefore improving the development cycle of the system, more particularly its design phase.

KNAML: A Knowledge Representation Language for Distributed Reasoning

“…At runtime, these goals are evaluated-using the Meta Agent reasoning process where appropriate-and the results are used to determine the path taken by the workflow. That multi-agent systems would benefit from a well-defined agent interaction protocol is clear [6]. The workflow agent orchestrates the behavior of the multi-agent system, and it does so in an architecturally neutral manner.…”