Energy saving in OFDMA cellular networks with multi-objective optimization

In this work, we develop a multi-objective optimization framework for reducing network wide energy consumption through base station (BS) sleep mode (SLM) in an OFDMA based cellular systems. While putting BSs in SLM, we take into account the effects such as coverage and overlap. After putting BSs in SLM, in order to ensure sufficient coverage, overlap (for hand-overs), and to achieve maximum network throughput of given set of active BSs, we consider the following objectives: power consumption minimization, area spectral efficiency maximization, coverage maximization, and overlap minimization; to find an optimal set of active BSs that are needed to support the required traffic demand. The system model incorporates accurate modeling of inter-cell interference considering pathloss, fading effects (shadowing and small scale fading), and activity of neighbouring cells. We provide analytical models to estimate cell capacity, network coverage, and blocking probability etc. from the distribution of signal-to-interference-plus-noise-ratio. Since, finding the optimal set of active BSs for SLM operation is a challenging combinatorial optimization problem, we use Genetic Algorithm to obtain the solutions of considered multi-objective optimization problem with fast convergence rate. We also study the trade-offs between the objectives. The results show that using the proposed mechanism significant amount of energy saving can be achieved while maintaining trade-offs between the objectives.