多Agent自动谈判情感决策模型研究

在基于多Agent的自动谈判中,谈判Agent往往采取一个相对固定的策略,使得Agent对动态变化的谈判环境缺乏适应性。根据情感理论的研究成果,本文提出建立情感决策模型来提高Agent的智能程度和适应性。在对谈判Agent情感决策过程进行分析后,建立了一般情况下的情感产生模型和决策模型,并通过在单属性自动谈判环境下,运用具体的情感决策模型和实际数据进行试验,实证了数据模型的有效性。本文的研究成果,可以应用到各种形式的基于多Agent的自动谈判中。

Research on Emotional Decision Model of Multi-Agent Negotiation

During the negotiation process of automated negotiation based on intelligent agent, the anegotiation agent usully adopts a fixed negotiation strategy, which limited its adaptability in the complex negotiation environment. Based on achievement of emotion theory research, we point out an emotional decision model to improve the degree of intelligence and adaptability of the agent. First, we analyse the emotional decision process of the agent in an automated negotiation, then propose an emotional generation model and emotional decision model in general. Taking single attribute negotiation as an example, we establish a concrete emotional decision model of the negotiation agent. In comparison with non-emotional model, it shows that our emotional decision model can get better result. The results of this research can be applied to various forms of automatic negotiation based on multi-agent.