Offer with Choices and Accept with Delay: A Win-Win Strategy Model for Agent Based Automated Negotiation

This paper formalizes a model for strategic negotiation by an automated agent in bilateral agentto- human multi-issue negotiations. Integrating insights from psychological and behavioral research, we hypothesize that compared to basic concession-based sequential-single offer and threshold-based immediate acceptance, a strategy based on simultaneous-equivalent offers and delayed acceptance makes a significant positive impact on both the economic and socialpsychological outcomes of negotiations. We tested these hypotheses using a 2x2 experimental study conducted with 110 industrial subjects who negotiated with an automated agent on a 4-issue business purchase task. MANCOVA results suggested that compared to a baseline condition when the agent did not employ the proposed strategy, settlement efficiency and buyer’s subjective responses were significantly higher when the seller agent provided simultaneous-equivalent offers, while the distance to Nash solution was significantly enhanced when the agent employed strategic acceptance delay. The findings confirmed the effectiveness of the proposed strategy and validated the efficacy of a novel implementation of an automated negotiation agent artifact