Two-Timescale Mobile User Association and Hybrid Generator On/Off Control for Green Cellular Networks With Energy Storage

In this paper, we study the energy-efficient (EE) cost minimization problem of green cellular networks, where base stations (BSs) are jointly powered by energy storages, i.e., batteries, external power grid with random electricity price, and local hybrid generators considering independent on/off strategies in long period and turn-on/off cost. The power consumption of each BS is correlated with the number of associated mobile users (MUs) and their random mobility. We formulate a stochastic optimization problem and decompose it into two subproblems with limited optimality gap. We solve the first MU association subproblem by minimizing the price-weighted power consumption of BSs under low-complexity matching algorithm. For the second energy control subproblem, we propose a novel two-timescale approach based on stochastic network optimization to tackle with the time-coupling turn-on/off cost and battery dynamic constraints. We derive the long-term time-averaged turn-on/off cost minimized by our approach including its upper bound and approximate expression. Lastly, extensive simulation results verify our theoretical performance analyses and show that our proposed algorithm can outperform the other three online benchmark algorithms.