Negotiation Planning: An AI Approach

Abstract

This paper develops a general approach for structuring and modeling negotiations, based on integration of Artificial Intelligence (AI) techniques with decision theoretic methods. Reasoning from past cases similar to the current negotiation, case-based reasoning, is combined with multi-attribute utility theory. The model also includes reasoning to accomodate departures from rationality in the behavior of the agents, and mechanisms to support dynamic problem restructuring. The model has been implemented in a computer program, the Persuader, that functions as a mediator in simulated labor negotiations. The main emphasis of the paper is in presenting the case-based reasoning aspect of the model.