A Cache-Aided Time-Domain Power Allocation for High-Speed Railway Communications

Deen Chen,Ke Xiong,Wanle Zhang,Bo Ai,Pingyi Fan,Khaled Ben Letaief
DOI: https://doi.org/10.1109/icc45855.2022.9838881
2022-01-01
Abstract:This paper investigates the cache-assisted power allocation in time domain for high-speed railway communications (HSRC), where train users are divided into real-time users (RUs) and non-real-time users (NRUs) from the time-sensitive perspective. RU’s real-time data rate demand is ensured by power allocation, and NRU’s data amount requirement is guaranteed by releasing cached content. In order to maximize the mobile service amount (MSA) of HSRC, an optimization problem is formulated to find the optimal cache switching time and power distribution under the constraints of total available energy, maximum power, caching and releasing causality, RU’s minimal data rate requirement and NRU’s minimal data amount requirement. Since the formulated problem is non-convex, a two-stage algorithm is proposed. In the first stage, we fix the cache switching time and transform the problem to be convex, and then use Karush-Kuhn-Tucker (KKT) condition to determine the optimal power distribution. In the second stage, one-dimensional search is employed to find the optimal cache switching time. Simulation results show that our proposed method is able to guarantee the RU’s data rate threshold all the time by sacrificing some MSA. Moreover, the increase of data rate threshold and speed lead to a decrease of MSA, while the cache usage rate has relatively weak influence on MSA.
What problem does this paper attempt to address?