A set cover based algorithm for Cell Switch-Off with different cell sorting criteria

The traffic distribution in cellular networks fluctuates in both time and space. This fluctuation results in some base stations (cells) being underutilized in light traffic conditions. Despite being underutilized, these cells still consume substantial amount of their energy. One possible technique to preserve this wasted energy is implementing the Cell Switch-Off (CSO) approach. In this approach, a common practice is to switch off cells based on their current loads. However, not only the cell load affects the switch-off procedure but also the order in which the cells are switched off (cell sorting). Hence, in this paper, we investigated different cell sorting criteria. The results illustrated that more energy can be preserved when sorting cells based on the number of users they can serve compared with the case of sorting cells based on their current load. To implement the CSO approach, we proposed a centralized greedy-add algorithm devised from the well known set cover problem. Simulation results showed that our algorithm outperformed the benchmark algorithm when the number of users per cell is large. The two algorithms were compared using the Urban-Micro (UMi) evaluation scenario.