摘要
针对Agent的劝说提议产生的研究不足导致谈判速度及拟人化程度都不高的现状,首先引人社会情感思想,考虑Agent的学习能力,提出了基于Agent的社会情感学习及其强度值的定义,并依据韦伯一费希纳定律,设计了基于Agent的社会情感学习算法;其次,进一步引人基于Agent的社会情感学习强度阈值的概念,提出了相应的行为决策模型,并给出了Agent在进行相应行为决策后的个体适应度值的计算方法;再次,引入中介Agent,结合差分进化算法和多属性效用理论,提出了相应的适应度函数,从而构建了基于Agent的社会情感学习及其行为决策的劝说提议产生模型;最后,通过算例仿真得出了采用该模型能使谈判速度及拟人化程度得到较大提高的结论。
In view of the slow negotiation speed and the low degree of personification caused by the lack of research on the generation of persuasive proposals,this paper firstly introduces the social emotion and the learning ability of agent and proposes the definition of agent-based social emotion learning and its intensity value in consideration of the learning ability of agent.And the agent-based social emotion learning algorithm is then designed according to the Weber-Fechner Law.Secondly,by further introducing the concept of the threshold of the agent-based social emotion learning intensity,it puts forward a corresponding behavioral decision model and the calculation method of individual fitness value of the agent after making behavioral decisions.Thirdly,via the introduction of an intermediary a gent,afitness function is proposed by combining the differential evolution algorithm and the multi-attribute utility theory,therefore the persuasion proposal generation model of the agent-based social emotion learning and behavioral decision-making is built accordingly.Finally,it reaches a conclusion that the proposed model can improve the negotiation speed and the personification degree through example simulation.
作者
伍京华
陈虹羽
许陈颖
陈秀兰
WU Jing-hua;CHEN Hong-yu;XU Chen-ying(School of Management.China University of Mining and Technology,Beijing 100083)
出处
《软科学》
CSSCI
北大核心
2020年第7期48-54,共7页
Soft Science
基金
国家自然科学基金项目(71972177)
中央高校基本科研业务费专项资金项目(2009QG03)。
关键词
差分进化算法
社会情感学习
行为决策
劝说提议
产生模型
differential evolution algorithm
social emotion learning
behavioral decision-making
persuasion proposal
generation model
