TY - GEN
T1 - Resource Allocation and Cluster Formation for Imperfect NOMA in DL/UL Decoupled HetNets
AU - Celik, Abdulkadir
AU - Radaydeh, Redha Mahmoud Mesleh
AU - Al-Qahtani, Fawaz S.
AU - El-Malek, Ahmed H.Abd
AU - Alouini, Mohamed-Slim
N1 - KAUST Repository Item: Exported on 2020-10-01
PY - 2018/1/25
Y1 - 2018/1/25
N2 - Being capable of serving multiple users with the same radio resource, non-orthogonal multiple access (NOMA) can provide desirable performance enhancements in a fair and spectral efficient manner. In this paper, we investigate the resource allocation (RA) and cluster formation (CF) aspects of NOMA for downlink (DL) uplink (UL) decoupled (DUDe) heterogeneous networks (HetNets). A non-ideal NOMA scheme is considered with power disparity and sensitivity constraints (PDSCs), delay tolerance, and residual interference after cancellation. Taking the PDSCs into account, we analytically show that using the DL decoding order limits UL-NOMA performance by that of OMA, while employing an inverse order result in a performance gain that is mainly determined by the channel gain disparity of users. Thereafter, a generic CF method is proposed for any type of user graph, which iteratively forms clusters using Blossom algorithm. Finally, highly non-convex RA problem is converted into a convex form by employing geometric programming (GP) where power and bandwidth are optimized to maximize network sumrate and max-min fairness objectives.
AB - Being capable of serving multiple users with the same radio resource, non-orthogonal multiple access (NOMA) can provide desirable performance enhancements in a fair and spectral efficient manner. In this paper, we investigate the resource allocation (RA) and cluster formation (CF) aspects of NOMA for downlink (DL) uplink (UL) decoupled (DUDe) heterogeneous networks (HetNets). A non-ideal NOMA scheme is considered with power disparity and sensitivity constraints (PDSCs), delay tolerance, and residual interference after cancellation. Taking the PDSCs into account, we analytically show that using the DL decoding order limits UL-NOMA performance by that of OMA, while employing an inverse order result in a performance gain that is mainly determined by the channel gain disparity of users. Thereafter, a generic CF method is proposed for any type of user graph, which iteratively forms clusters using Blossom algorithm. Finally, highly non-convex RA problem is converted into a convex form by employing geometric programming (GP) where power and bandwidth are optimized to maximize network sumrate and max-min fairness objectives.
UR - http://hdl.handle.net/10754/623244
UR - https://ieeexplore.ieee.org/document/8269139
UR - http://www.scopus.com/inward/record.url?scp=85046368423&partnerID=8YFLogxK
U2 - 10.1109/GLOCOMW.2017.8269139
DO - 10.1109/GLOCOMW.2017.8269139
M3 - Conference contribution
AN - SCOPUS:85046368423
SN - 9781538639207
SP - 1
EP - 6
BT - 2017 IEEE Globecom Workshops (GC Wkshps)
PB - Institute of Electrical and Electronics Engineers (IEEE)
ER -