Intelligent Energy and Traffic Coordination for Green Cellular Networks With Hybrid Energy Supply

In energy-harvesting-enabled networks, the intermittent and randomly distributed renewable energy imposes severe challenges in reliably supplying the time-varying mobile traffic. To tackle this issue, we reshape the spatial renewable energy and mobile traffic by exploiting the approach of energy sharing and load shifting, with the objective of minimizing the grid energy expenditure of cellular networks powered by both grid and renewable energy. We formulate this problem as a mixed-integer nonlinear programming, which is proved to be NP-hard. For centralized networks, we first devise a cost-efficient centralized algorithm leveraging the univariate search technique, which can find the near-optimal solutions with the advantages of low complexity and fast convergence. Specifically, by jointly optimizing the spatial distribution of renewable energy and mobile traffic, the centralized algorithm achieves a good match between the renewable energy supply and the total energy demand at each base station (BS), such that the grid energy expenditure of the whole network is greatly reduced. For distributed networks, we further propose a three-phase distributed control policy in which BSs and mobile users adjust their strategies independently only with their local information. Finally, we present extensive simulations to investigate the convergence and effectiveness of our proposed algorithms and demonstrate the achieved energy conservation gain compared with the existing schemes.