Multiple Agent Coopoerative Design Evaluation Using Negotiation

Abstract

This paper presents a form of multiagent cooperative problem solving where computational agents, paralleling human design, manufacturing and assembly participants, evaluate a design and comment on it from their unique expert perspectives. The evaluation process utilizes a novel form of incremental negotiation called knowledge-based negotiation where agent conflicts are resolved through the use of shared knowledge representations called shareable perspectives. Two agents’ shareable perspectives can be linked together to form an interagent issue relation which are grouped into a relational network and maintained by a third-party arbitrator agent. The arbitrator uses the relations to develop alternatives during conflict resolution between agents performing a design review. Through the use of arbitrator suggested shareable perspectives, a form of issue unlinking occurs and this allows the agents to consider additional viable alternatives which they might have dismissed earlier. This behavior is similar to integrative bargaining found in human negotiation where conflicting parties are enlightened about secondary benefits of a particular proposal. The resulting research is demonstrated as an implementation within a knowledge-based tool called the Designer Fabricator Interpreter (DFI) System.