Joint Design of Content Update, Push and Delivery Based on Noma and Swipt in Wireless Caching Networks

This paper proposes a novel content-centric framework for joint content update, push, and delivery in wireless caching networks. We employ a two-stage transmission strategy by grouping users based on their distribution locations and content caching status of the content server (CS). Users can be served either by the base station (BS) through multicast transmission or by the CS covering them and caching the requested contents. The content update scheme is designed based on the caching function that takes the request popularity and file size into account. Afterwards, the content push and delivery process is conducted in two stages. In the first stage, the requested and updated contents for multicast groups and the CS are delivered, while the CS harvests energy by employing the function of simultaneous wireless information and power transfer (SWIPT). In the second stage, the CS serves its covering users utilizing non-orthogonal multiple access (NOMA) transmission. Based on the proposed caching and transmission designs, we formulate the sum rate optimization problem, which is challenging due to its non-convexity. Thereafter, we convert it into a convex second-order cone programming problem using semi-definite relaxation (SDR), successive convex approximation (SCA), and other approximation techniques, and solve it with CVX solvers. Simulation results exhibit the fast convergence property of the proposed algorithm and demonstrate the advantages of our twostage content transmission approach, particularly in terms of the content hit rate and the sum rate.