A Decision Support Algorithm for Ontology-Based Negotiation Agents within Virtual Enterprises

This paper describes an application of a distributed knowledge-based multi-agent architecture to support automated negotiation between and within virtual enterprises (VE). Considering constraints of traditional negotiation approaches (predetermined protocols as well as hard-coded and non-adaptive strategies), we combine the agent paradigm with ontology-based solutions to overcome these limitations. Our negotiation ontology, used by agents, acts as a general framework that defines the basic terminology, interaction and protocols enabling agent to understand each other and to reach an agreement. Our multi-agent architecture consists of two types of agents that have rule-based behaviors. One is the Auctioneer that starts, monitors and ends an auction process in the name of a user that offers a particular product/service. Every new user who joins an auction has a configured MarketAgent that takes a participant role bidding for a product/service. To help an agent in its decision, a special decision support algorithm is implemented to recommend the best negotiation strategy, based on the user interests. We also describe the initial implementation of the system using JADE and JESS.