Agent negotiation as fuzzy constraint processing

This paper presents a general problem-solving framework for the modeling of agent negotiation via fuzzy constraint processing. In this approach, fuzzy constraint processing serves to improve the agent´s individual negotiation strategy, and also to represent the agent negotiation mechanism and interaction framework for multi-objective agent negotiation. We formulate agent negotiation as a distributed fuzzy constraint satisfaction problem (DFCSP). Fuzzy constraints, in this way, can be used not only to define the needs of agents together with their priorities, urgency, and preferences, but also the similarity between the offers and counter-offers. On the other hand, basing on fuzzy constraint-based problem-solving, we propose concession strategy for fuzzy constraint relaxation, and trade-off strategy for modifying the current solution. The purpose is to generate an acceptable solution with a maximal satisfaction degree. Additionally, we also define a state transition function which specifies the operational semantics for representing the temporally changing states and agent interaction. To demonstrate the effectiveness of this approach, an example of badge-trading is provided