Resource Allocation in STAR-RIS-Aided Networks: OMA and NOMA

Simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) is a promising technology to achieve full-space coverage by splitting the incident signal into the transmitted and reflected signals towards both sides of the surface. This paper investigates the resource allocation problem in a STAR-RIS-assisted multi-carrier communication network. To maximize the system sum-rate, a joint optimization problem of channel assignment, power allocation, and transmissionand reflectionbeamforming at the STAR-RIS for orthogonal multiple access (OMA) is first formulated, which is a mixed-integer non-linear programming problem. To solve this challenging problem, we first propose a channel assignment scheme utilizing matching theory and then invoke the alternating optimization-based method to optimize the resource allocation policy and beamforming vectors iteratively. Furthermore, the sum-rate maximization problem for non-orthogonal multiple access (NOMA) with flexible decoding orders is investigated. To efficiently solve it, we first propose a location-based matching algorithm to determine the sub-channel assignment, where a transmitted user and a reflected user are grouped on a sub-channel. Based on this transmission-and-reflection sub-channel assignment strategy, a threestep approach is proposed, where the decoding orders, beamforming-coefficient vectors, and power allocation are optimized by employing semidefinite programming, convex upper bound approximation, and geometry programming, respectively. Numerical results unveil that: 1) For OMA, a general design that includes same-side user-pairing for channel assignment is preferable, while for NOMA, the proposed transmission-and-reflection scheme can achieve comparable performance as the exhaustive search-based algorithm. 2) The STAR-RIS-aided NOMA network significantly outperforms the networks employing conventional RISs and OMA. C. Wu and X. Gu are with the School of Electronic and Information Engineering, Harbin Institute of Technology (HIT), Harbin, 150001, China. (e-mail: {wuchenyu, guxuemai}@hit.edu.cn). X. Mu is with the School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, 100876, China (email: muxidong@bupt.edu.cn). Y. Liu is with the School of Electronic Engineering and Computer Science, Queen Mary University of London, London E1 4NS, U.K. (email: yuanwei.liu@qmul.ac.uk). X. Wang is with the Department of Electrical and Computer Engineering, Western University, London, ON N6A 5B9, Canada (e-mail: xianbin.wang@uwo.ca).