Dianxin kexue (Jul 2024)
IRS-assisted D2D system resource allocation and RCG beamforming optimization in H-CRAN
Abstract
Intelligent reflecting surface (IRS)-assisted device-to-device (D2D) communications in heterogeneous cloud radio access network (H-CRAN) were investigated as research background. Resource allocation and Riemannian conjugate gradient (RCG) beamforming optimization were studied, with the objective of system sum rate maximization. System sum rate was formulated as the optimization objective, which was subject to several constraint conditions such as sub-channel reuse coefficient, transmit power threshold, as well as the modulus of the IRS reflection coefficient. To solve the formulated mixed-integer non-linear programming problem, a channel-strength-based deferred acceptance algorithm was proposed to obtain channel reuse indicators. The problem was then decomposed into two subproblems. For transmit power optimization subproblem, successive convex approximation (SCA) was used to solve it. For IRS beamforming optimization subproblem, the beamforming vector constraint was transformed into a complex circular manifold and Riemannian conjugate gradient (RCG) algorithm was implemented to solve it. Simulation results show that, when IRS reflecting elements is 50 and base station maximum transmit power is 46 dBm, compared with the existing channel allocation scheme and random channel allocation scheme, the proposed scheme enhances sum rate performance 5.2 bit/(s·Hz) and 14.6 bit/(s·Hz) respectively. Compared with the communication scenario without IRS, sum rate performance significantly promotes nearly 31.2 bit/(s·Hz) with the deployment of IRS.
