From: Towards intelligent user clustering techniques for non-orthogonal multiple access: a survey
Clustering scheme | Related work | Scenario | Assumptions | Objective functions | Application | Findings |
---|---|---|---|---|---|---|
Mathematical optimization | Downlink | CSI is known at base station | Sum rate maximization | MISO | (1) It performs better than random clustering (2) Jointly user clustering and power allocation | |
Game-theoretic approach | [9] | Downlink | The channels assumed to be Rician fading | Maximize the energy efficiency | MIMO | Joint user clustering, passive beamforming and power allocation |
Matching algorithm | [10] | Downlink | The channels assumed to be Rician fading | Maximize the energy efficiency | RIS assisted NOMA | Joint user clustering, passive beamforming and power allocation |
Other user clustering techniques | [12] | Downlink | Users inside a cell are independent of path loss effect | Sum rate maximization | MU-MIMO | To suppress multi-user and inter-symbol interference |
 | Downlink | Assume no inter-cluster interference | Higher spectral efficiency | Massive MIMO | To improve performance by NOMA beamspace with efficient user clustering | |
 | [15] | Downlink | Location information is assumed to be known by the UE | Sum rate maximization | 5G mm-wave | Location aided clustering approach is used for cluster assignment and improves the performance as compared to other beamforming schemes |
 | [16] | Uplink | BS knows perfectly all the channel gains | Reduce system latency | MISO | To solve the problem of user clustering by optimally resource block allocation with time-based proportional fairness |
 | [17] | Downlink | CSI is known at the base station, fixed user distances | Improves transmission power | MISO | Iterative power and joint user clustering algorithm reduces the system’s power consumption |
 | [19] | Downlink | CSI is known at the base station | Sum rate maximization | SISO | This schemes used Brute Force Search method and improves system performance |
 | [20] | Uplink | No intercell interference from other neighbouring cells | Sum rate maximization | Narrow band IoT | (1) Joint user clustering and resource allocation of MTC devices (2) Satisfying the QoS requirements and transmission power in each cluster |
 | [21] | Uplink | CSI is known at the base station | Higher spectral efficiency | MISO | Provides optimal solution for decoding order and power control |
 | [11] | Downlink | CSI is known at the base station | Minimize total power consumption | SISO | (1) Consider decoding power consumption (2) Compressive sensing method is used to solve the convex problem |