Residual Energy-Aware Caching in Energy Harvesting-Based Mobile D2D Network

Abstract

Caching popular contents on mobile devices is a promising technique to alleviate the backhaul data rate requirements. Since both file placement and data exchange among mobile devices consume energy, the energy status of the device can have a significant impact on the caching utility of the whole system. This work considers the caching optimization in a cellular network, where mobile devices are charged with energy harvested from the ambient environment. As the devices can collect the file segments from the local storage via device-to-device (D2D) links and cellular links, we aim at minimizing the percentage of file segment that should be collected from the base station (BS) by optimizing the file placement scheme on the devices. We prove that the optimal file placement scheme cannot be found in polynomial time by showing the intractability of the problem. Since the problem is NP-hard, we propose a residual energy-aware file placement algorithm based on the popularity distribution of contents and causality of energy arrival. In particular, we show that the proposed algorithm can fastly find a suboptimal file placement solution using dynamic programming. Extensive simulation results show that the proposed algorithm can effectively reduce the percentage of file segments collected from the BS compared with other two conventional caching methods.