Distributed Inter-BS Cooperation Aided Energy Efficient Load Balancing for Cellular Networks

Abstract

We propose a distributed cooperative framework among base stations (BS) with load balancing (dubbed as inter-BS for simplicity) for improving energy efficiency of OFDMA-based cellular access networks. Proposed inter-BS cooperation is formulated following the principle of ecological self-organization. Based on the network traffic, BSs mutually cooperate for distributing traffic among themselves and thus, the number of active BSs is dynamically adjusted for energy savings. For reducing the number of inter-BS communications, a three-step measure is taken by using estimated load factor (LF), initializing the algorithm with only the active BSs and differentiating neighboring BSs according to their operating modes for distributing traffic. An exponentially weighted moving average (EWMA)-based technique is proposed for estimating the LF in advance based on the historical data. Various selection schemes for finding the best BSs to distribute traffic are also explored. Furthermore, we present an analytical formulation for modeling the dynamic switching of BSs. A thorough investigation under a wide range of network settings is carried out in the context of an LTE system. Results demonstrate a significant enhancement in network energy efficiency yielding a much higher savings than the compared schemes. Moreover, frequency of inter-BS correspondences can be reduced by over 80%.