Affective Computing have become a new aspect of development in the field of Artificial Intelligence, it makes great sense to the implement of harmonious interaction between human and computers and the management of intelligent information. However, Affective Computing is confronted with a lot of problems and challenges now.Granular Computing is a new method to dealing with the information and knowledge in the field of Artificial Intelligence, nowadays, more and more researchers use it in different fields.At the basic of Artificial Intelligence and Artificial Mentality, aim at how to implement affective comprehending,affective computing and affective creating, with the needs of affective modeling,this paper using the theory of Granular Computing to explore,research and implement a model of Affective Computing based on Information Granular.In the first place, this paper establishes a model based on Multi-Agent. In this model, we consider the emotion is made up of the basic emotion,faith,desire,intent and programming, moreover, at the basic of the relationship of "cognize" and "emotion", we import the sense reasoning agent and emotion reasoning agent to this model. And then, we use some mathematic method to establish an emotion repository. Based on this emotion repository, we use the method of Information Granular data reduction to describe the process of emotion decision-making. The model has been realized, experiment results show that it is efficient to transact simple emotions. |