Proactive Self-Healing Analysis-Framework Based on Discrete-Time Markov Decision Process in 5G Network and Beyond

摘要

Future cellular network especially 5G needs to be efficient, reliable and agile to meet the targeted performance requirements. Considering the density and complexity anticipated for 5G network, more intelligence in all layers of the network management is required. Self-Organizing Network (SON) is one of the promising technologies in the area of cellular network management. Self-Healing (SH), self-optimization and self-configuration are three well-known categories which encompass almost all autonomous techniques in SON. In this paper, we develop a framework for analyzing proactive SH in contrast to reactive SH taking into account the effect of recovery actions accomplished in sub-health states. Our framework is adapted to the Markov Decision Process (MDP) and uses linear programing (LP) to determine the optimal action or policy maximizing an operator-defined performance metric. Numerical results representing different types of scenarios with diverse targets demonstrate the applicability of the proposed analytical framework.