A Generic Framework for Automated Multi-attribute Negotiation

Agents in a negotiation may have to negotiate multiple issues simultaneously. Automated multi-attribute negotiation provides an important mechanism for distributed decision makers to reach agreements on multiple issues. Moreover, it also furnishes the opportunity to reach "win-win" solutions. In this paper, we first provide a survey that synthesizes the research on multi-attribute negotiation. We discuss the limitations of the existing research and conclude that three key issues need further study: incomplete information, Pareto optimality, and tractability. We then present a generic framework for automated multi-attribute negotiation with two new mechanisms that address the above issues. Finally, we discuss the challenges and directions for future work.