Minimizing Energy Cost of Base Stations with Consideration of Switching on/off Cost

Abstract

Base stations (BSs) are densely deployed in cellular networks to meet the increasing peak demand of mobile data traffics. However, since the distribution and traffic demand of user equipment (UE) fluctuate in time and over space, a number of BSs are underutilized in certain period of time. Therefore, switching on/off BSs dynamically according to the variations of UE distribution is regarded as a promising way to save energy in cellular networks. The existing works on BS switching do not consider the energy cost for switching on/off BSs, and they take the UE distribution and network traffic in a time period as given in advance. On the contrary, we consider the problem of minimizing energy cost of BSs with switching cost included. To solve this problem, we propose an online BS switching algorithm along with a lightweight UE distribution prediction strategy. Simulations are carried out in an example network to show that our online scheme outperforms the existing methods and its performance is close to the offline algorithm, which relies on the accurate network information in the whole time period.