A stage model for agent-based emotional persuasion with an adaptive target: From a social exchange perspective

Abstract

Agent-based automated negotiation is a promising approach for business negotiation; however, most studies in this field focus on negotiation strategies, ignoring the role of similarities between negotiating parties and the interactive social nature of negotiations. This study considers similarities between proposed and expected values in automatic negotiations to improve the agent perception ability. A cloud model was innovatively applied to calculate similarities, and overcome subjective characterization of proposed values in negotiations. Additionally, two social factors, i.e., emotion and familiarity, were identified in negotiations, the Web-Fechner law was applied to model an agent’s generation of emotion, and a familiarity function was constructed to measure the level of familiarity between negotiating agents. Finally, a stage model for agent-based emotional negotiation which integrates similarity, emotion, and familiarity was established. A series of numerical experiments were conducted and comparative analyses of the proposed model with other similarity measures and existing negotiation models were conducted to test the performance of the proposed model. The results show that the proposed model can lead to more successful negotiations and fairer negotiating outcomes and simultaneously, maintain competitive negotiation efficiency.