An Approach to One-to-Many Concurrent Negotiation

Abstract

Automated negotiation is a market mechanism that can improve the electronic transactions in terms of both the time needed to complete the transactions and the quality of the transactions. The quality of a transaction can be measured—for example—by its utility. This paper addresses the problem of the bidding strategy for a buyer agent negotiating concurrently with multiple seller agents over multiple distinct objects, e.g., resources. Each distinct object has multiple providers and a single negotiation issue. In our scenario, we assume that all objects have the same negotiation issue. For the stated scenario, we propose novel coordination mechanisms that manage the bidding strategy for the buyer agent benefiting from the different objectives of the seller agents over all negotiation objects and from the different objectives of the seller agents of a given object. We propose three dynamic bidding strategies: global, local and hybrid. The strategies are tested against two other benchmark negotiation strategies used in the literature. The experimental results show that the proposed strategies perform well when compared with the other benchmark strategies under tough negotiation conditions.