Satellite-Terrestrial Assisted Multi-Tier Computing Networks with MIMO Precoding and Computation Optimization

Changfeng Ding,Jun-Bo Wang,Yijian Chen,Hongkang Yu,Ming Cheng,Min Lin,Jiangzhou Wang

DOI: https://doi.org/10.1109/twc.2023.3311104

IF: 10.4

2023-01-01

IEEE Transactions on Wireless Communications

Abstract:In this paper, satellite-terrestrial assisted multi-tier computing networks (STMTCN) are proposed to satisfy the growing computation demands of user terminals (UTs) in next generation wireless networks. In the STMTCN, UT’s computation task can be processed at different computing entities and a multi-tier computation model named computing depth is proposed to better reflect the multi-tier computing process. Then, we formulate a weighted sum energy consumption minimization problem via jointly optimizing UT-satellite association, computing depth, multiple-input multiple-out (MIMO) precoding, and computation resource allocation. The non-convex optimization problem is decomposed into four subproblems, each of which is solved iteratively. Specifically, the UT-satellite association subproblem is solved by quadratic transform based fractional programming and Lagrangian dual method and a closed-form expression is obtained. The computing depth for local tier and the satellite tier is solved respectively with first-order Taylor expansion. Then, MIMO precoding subproblem for UT and satellite offloading is solved by quadratic transform and interior point method (IPM). Finally, the computation resource allocation for UT and satellite is obtained in a closed-form expression and the GW computation resource allocation is solved by using IPM. Simulation results show that the proposed STMTCN and algorithms can fulfill the UT’s computing demands with low energy consumption.

telecommunications,engineering, electrical & electronic