Self-Configuration in Autonomic Systems Using Clustered CBR Approach

Abstract

Self-configuration is one of the key properties of autonomic systems. We apply an experience-based artificial intelligence approach known as case-based reasoning (CBR) in order to help autonomic manager to devise new configuration solution. Searching the entire case-base on occurrences of every new problem is a time consuming task. We propose to cluster the case-base and classify each new problem among one of the clusters. Our approach to reduce the search space promises to achieve efficiency as well as accuracy. We performed experiments on a simulation of autonomic forest fire application and achieved inspiring results.