A Low Complexity Joint User Grouping and Resource Allocation Algorithm in Massive MIMO Systems

In the downlink user grouping of multiuser massive MIMO systems, the difference of packet sizes between users assigned to the same group may cause a waste of resource block (RB), which deteriorates the overall system throughput. In this paper, an algorithm with rather low complexity is proposed combining user grouping with resource allocation. Specifically, we group users in terms of each RB. In the process of user grouping on each RB, we obtain a candidate set which is much smaller than the original user set according to the user correlation to reduce the algorithm complexity. Users are assigned to the same group (i.e., RB) with objective to reducing multiuser interference. The user grouping in a RB is finished after this RB is completely allocated to its users. The proposed algorithm ensures that users with larger packets can be assigned to more groups (i.e., more RBs) while users with smaller packets are assigned to less groups. Finally, compared with other existed algorithms, the advantages of the proposed algorithm in system throughput and fairness are validated by simulations.