Energy saving in heterogeneous cellular network via transfer reinforcement learning based policy

Abstract

Energy efficient operation of heterogeneous networks (HetNets) has become extremely crucial owing to their fast increasing deployment. This work presents a novel approach in which an actor-critic (AC) reinforcement learning (RL) framework is used to enable traffic based ON/OFF switching of base stations (BSs) in a HetNet leading to a reduction in overall energy consumption. Further, previously estimated traffic statistics is exploited in future scenarios which speeds up the learning process and provide additional improvement in energy saving. The presented scheme leads to up to 82% drop in energy consumption which is a quite significant amount. Furthermore, the analysis of system delay and energy saving trade-off is done.