Agenda-Based Automated Negotiation Through Utility Decomposition

Zongcan Li, Rafik Hadfi, Takayuki Ito

Abstract: The success of a negotiation depends mainly on the strategies of the negotiators and the problem domain. It is common for negotiators to rely on an agenda to simplify the process and reach better deals. This is particularly true when the negotiators’ preferences are defined over multiple issues. Using an agenda to explore and decompose the interdependencies between the issues is one way to address this problem. This paper applies the classical divide-and-conquer approach to automated negotiations through utility decomposition and bottom-up agenda construction. The approach does not impose an agenda from the top level of the negotiations but builds it bottom-up, given the individual utility functions of the agents and the relationships between the issues. We implemented our method in a novel protocol called the Decomposable Alternating Offers Protocol (DAOP). The protocol reduces the cost of exploring the utility spaces of the agents and the generation of optimal bids. As a result, the divide-and-conquer algorithm positively influences the global performance of an automated negotiation system.