Dynamic Programming Principle for Automatic Negotiations

Abstract

In automatic negotiation, an automated agent is trained to negotiate on behalf of a human negotiator. We consider that the domain of negotiations is known to both the agent and its opponents. In this setting, a greedy concession algorithm (GCA) is employed to find an optimal policy for the agent when the agent's belief about the opponent is known. However, the GCA is computationally expensive to certain class of policies. In this article, we propose a reverse greedy concession algorithm that is computationally less expensive for such class of policies. We also provide an alternate proof to establish that the GCA is a dynamic programming principle.