Graphical modelling approach for monitoring and management of telecommunication networks

摘要: The recent focus on monitoring and managing telecommunication networks in a more efficient autonomic way has led to the widespread application of machine learning (ML) approaches for network management tasks. In order study behaviour evaluate performance such systems, it is requirement that suitable modelling framework exists. work presented here addresses this need by comparing existing ML-based proposing solution which employs prediction capabilities Bayesian (BN) approach. It also formulates BN-based decision support system providing real-time call admission control (CAC) decisions next generation (NGN) environment. provide realistic simulation environment, surveys computer simulators BN choose most test proposed models. novelty research validated through offline online evaluation networks-based (BNAC) terms metrics packet delay, loss, queue size blocking probability. This paper concludes BNAC approach appropriate choice implementing CAC autonomic.