Energy-efficient Base Station Control with Dynamic Clustering in Cellular Network

Due to increasing awareness of environmental and economic issues for network operators, energy-efficiency has received much attention lately and becomes one of the major design goals in wireless networks. Since base stations (BS) contribute most power consumption of entire cellular network, in this paper, we propose a novel power saving method, called clustering BS-off (CBSO) scheme by taking into account potential heterogeneous traffic distribution. In CBSO scheme, BSs are cooperated to form several clusters, which are changing dynamically and adaptively with respect to the variance of traffic load along space and time. Both centralized and distributed CBSO protocols are proposed to group BSs with similar traffic load. Instead of adopting a unified BS-off strategy in whole network, the proposed CBSO schemes run BS-off matching scheme in order to find the best BS-off mechanism for each cluster separately. Therefore, low traffic zones can perform more aggressive BS-off strategy than the hot spots so that further improvement on energy efficiency can be achieved. Experimental results show that the proposed CBSO algorithms can save over 50% aggregated network power consumption, and outperform the traditional unified BS-off strategy. Moreover, the impact of changing the transition cost on selecting BS on/off modes and the operation intervals is also evaluated.