Preference Based Automated Negotiation for Service Procurement in Multi-Agent Systems

This paper addresses the problem of automated agent-based one-to-many negotiation for multi-issue service procurement and presents a robust negotiation mechanism to enhance the procurement process. Software agents represent service sellers and service buyers. Agents are assumed to have goals and constraints when buying and selling services. Automating the service procurement process can provide fast and efficient procurement results. The presented negotiation mechanism uses a concession tactic that facilitates service procurement from the best provider amongst multiple prospective providers. The strategy of the tactic is based on offering more concessions on the negotiation issues that are believed to be more preferred to the opponents. In other words, the tactic considers the preferences of the opponent agents on different negotiation issues (attributes) before selecting the amount of the concession on each issue. The negotiation mechanism is validated experimentally.