Decision Making in Autonomic Managers Using Fuzzy Inference System

Abstract

Inspired from natural self-managing behavior of human body, autonomic systems promise to inject self-managing behavior in software systems. Such behavior enables self-configuration, self-healing, self-optimization and self-protection capabilities in software systems. Fuzzy Inference System (FIS) is a decision methodology suitable for the vague and imprecise application domains such as software systems. Building a complete crisp rule-based system is a hard job in such complex domains because of large number of enumerations of possible rules. In literature, FIS has been successfully applied in the domains of decision analysis, automatic control and expert systems. In this paper, we have proposed to apply FIS for diagnosis and planning purposes in the autonomic manager. We implemented the proposed architecture on a simulation of Autonomic Forest Fire Application (AFFA) and achieved up to 88% accuracy.