A new rational model of agent for autonomic computing

Autonomic computing is a computing system that can manage itself by self-configuration, self-healing, self-optimization and self-protection. With the nature of autonomy, reactivity, sociality, and pro-activity, software agents are promising to make autonomic system a reality. Traditional BDI model sets up a generic rational model of agent by establishing models of mental state of agent from three aspects, i.e., belief, desire and intention. However, in autonomic computing system, the emphasis is on the mutual relationships between agent and its environment. BDI model is too general and lack of pertinence. Furthermore, it also has logical omniscience and side-effect problems. This paper presents a state feeling-action (SFA) rational model of agents for autonomic computing. Based on possible world semantics, this model points out the corresponding relationships between consciousness and behaviors of autonomic element, and describes the interoperation between agents and their environment. The syntax and semantics of SFA model are defined in detail in this paper, and proved that logical omniscience problem is avoided in SFA model.