Emotion modeling in autonomic computing

Autonomic unit is the essential element and the most basic component of autonomic systems. Having the ability of handling emotions is considered to make autonomic unit more intelligent, more communicative and more social. In this article we try to build an emotion model for agents that act as autonomic units in autonomic systems and can self-manage its behavior to achieve the best performance. From four levels we define four models progressively and the top-level model is defined on the foundation of the three below, combining them as a unity. By given arousing axioms theoretically of several typical emotions, the model is proved to be effective for reasoning emotions of autonomic unit, and can be easily extended for dealing with other type of emotions. On this base an architecture of emotional autonomic unit is described. An example shows that emotion plays an important role in intelligent decision making, which proves the validity of the model and the special function in practice.