摘要： 基于Agent的情感勸說是在基于Agent的勸說式辯論談判的基礎上加入了情感成分,使得勸說更加符合人的思維,勸說結(jié)果更加符合實際。其中的合作主體選擇模型更為重要,而該模型的構(gòu)建建立在有效的合作主體評價模型基礎之上。針對現(xiàn)有研究不足,首先對基于Agent的情感勸說的情感進行分類,依據(jù)該分類提出了相應的情感產(chǎn)生的形式化模型,并結(jié)合心理學的情感第一定律,對Agent情感進行量化,提出了Agent情感強度的計算方法;其次,將基于Agent的情感勸說的評價歸納為對Agent主體的提議和信度的評價,隨后運用多屬性效用函數(shù)建立了相應的Agent主體提議評價模型,運用模糊隸屬度函數(shù)建立了相應的Agent主體信度評價模型;再次,在以上基礎上,將情感作為新的評價因素,建立了其中關于Agent主體情感的評價模型,并綜合提出的三個評價模型,構(gòu)建了相應的基于Agent的情感勸說的合作主體選擇模型;最后,通過算例仿真證明了該模型的合理性和有效性,得出了相應的管理啟示。 With the rapid development of the Internet and e-commerce, automatic negotiation has become a hot research topic. Due to its unique intelligence and interaction, agent can realize complex negotiation and has great advantages in the field of automatic negotiation. Agent-based debate negotiation can send information other than the proposal by making the negotiation more efficient. Persuasion debates consider more persuasion factors and make the negotiation process more rational. In order to adapt to the development of artificial intelligence efficiently, the emotion is introduced into the field of agent persuasion, which makes the agent's behavior more in line with human thinking and better simulates human emotional decisions. In consequence, the negotiation results are easier to understand and interpret. At this moment, agent has not only emotional ability, but also logical computing ability, which improves agent's adaptability in dynamic and complex environments, and achieves the goal of automatic negotiation effectively.In related research, different scholars and experts have studied agent-based persuasion and agent-based emotion from different perspectives. However, the following problems exist. First, the research on emotion is not comprehensive, especially the quantification of emotion. Second, the current research about the application of emotional theory in the agent persuasive field is not enough. Third, most of agent persuasion study is about interaction mechanism, and there is little research on the choice of cooperative subjects, especially the selection of cooperative subjects considering emotional factor. In the study of the selection of cooperative subjects, the selection model is more important, and the construction of the model is based on an effective cooperative subject evaluation model. Therefore, aiming at the shortcomings of the existing research, we classify the emotion-based emotion persuasion based on agent. According to the classification, we propose a corresponding formal model of emotion generation, and combine the first law of emotion intensity of psychology to quantify the emotion of agent. The agent's emotion intensity calculation method is proposed. Secondly, the evaluation of agent-based emotional persuasion is summarized as the evaluation of the agent's proposal and reliability. Subsequently, the multi-attribute utility function is used to establish the corresponding agent's proposal evaluation model, and the fuzzy membership function is used to establish the corresponding agent's reliability evaluation model. On the basis of the above two evaluation models, emotion is used as a new evaluation factor to establish an evaluation model of the subject's emotions. Three evaluation models are put forward comprehensively to construct a corresponding Agent-based emotional persuasive cooperation subject selection model. To illustrate the effectiveness of the above model, an example was designed and simulated. The simulation results show that emotion in emotional persuasion has an important influence on the choice of cooperative subjects. Among them, positive emotion enhances the comprehensive evaluation value in emotional persuasion and negative emotion weakens the comprehensive evaluation value. In the end, due to the influence of emotion, there are obvious differences between comprehensive evaluation values, which can significantly affect the ultimate selection of cooperative subjects. Based on this, the rationality and effectiveness of the model are verified. At the same time, it also illustrates the important influence of emotion on agent-based persuasiveness.
APP 