Automated rational recovery selection for self-healing in mobile networks

Self-healing is a key functionality of Self-Organizing Networks (SON). There have already been promising research results on Degradation Detection and Root Cause Diagnosis. However, the complex task of Recovery Selection, i.e., the determination of the best recovery action for an uncertain diagnosis of a degradation considering the operational goals of the network operator, is rarely investigated so far. In this paper, a two-step rational Recovery Selection system and its integration into a self-healing process for mobile networks is presented. Thereby, Recovery Selection first exploits technical recovery knowledge expressed in rules to determine the set of possible recovery actions for a degradation situation. Second, it determines the degree of rationality of these actions based on the operational goals of the operator. Therefore, it draws on decision theory in order to handle the uncertainty in the problem situation and the possibly conflicting preferences of the operator. The presented system enables automatic rational Recovery Selection which provides a baseline for self-healing in mobile networks.