TACT: A Transfer Actor-Critic Learning Framework for Energy Saving inCellular Radio Access Networks

Abstract

The explosive recognition of clever telephones and tablets has ignited a surging visitors load call for for radio access and has been incurring big electricity consumption. Currently, over 80% of the electricity intake takes place within the radio get entry to networks particularly the bottom stations (BSs). The purpose at the back of this is essentially due to that the existing BS deployment is on the idea of height site visitors loads and generally remains active regardless of the heavily dynamic visitors load versions. In this paper, we enlarge the studies over BS switching operations, which have to match up with visitors load versions. Instead of relying on the dynamic site visitors hundreds which are nevertheless pretty tough to precisely forecast, we firstly formulate the site visitors versions procedure. And we layout BS switching operation scheme to minimize the strength consumption of RANS.