Multiple Offer Strategy for Automated Negotiation Agents

Automated negotiation agents negoatite issues of an e-commerce transaction with human consumers on behalf of e-commerce vendors, and they can increase the financial benefits of the consumer and the vendor jointly. Both the artificial intelligence community and the economics (game-theory) community have proposed methods for negotiation agent design. The focus has been on agents that are restricted to make a single counteroffer to the consumer at every round of the negotiation. Recent studies from the psyhchology community, however, indicate that making multiple counteroffers per negotiation round can be beneficial to both the consumer and the vendor. In light of this, we design an automated software agent that makes multiple offers at every round. We devise a probabilistic strategy that guides the agent to find the optimal set of counteroffers.