Search

In this work, we propose an uplink power control (PC) framework compliant with the technical specifications of the fifth generation (5G) wireless networks. We apply the fundamentals of reinforcement learning (RL) to develop a power control algorithm able to learn a strategy that enhances the total data rate on the uplink channel and mitigates the neighbor cell interference. The base station (BS) uses a set of commands to specify by how much the user equipment (UE) transmit power should change. After implementing such commands, each UE reports a set of information to its serving BS, and this, in turn, predicts the next commands to achieve a suitable UE transmit power level. The BS converts the UE reports into rewards according to a predefined cost function, which impacts the longterm behavior of the UE transmit power. The simulation results indicate a near-optimum performance of the proposed framework in terms of total transmit power, total data rate, and network energy efficiency. It provides a self-exploratory power control strategy that overcomes soft dropping power control (SDPC) with similar signaling levels.

This work proposes a framework for multiuser massive Multiple Input Multiple Output (MIMO) systems which is composed of three parts – clustering, grouping, and scheduling – and aims at maximizing the total system data rate considering Quality of Service (QoS) constraints. We firstly use a clustering algorithm to create clusters of spatially correlated Mobile Stations (MSs). Secondly, in the grouping part, we select a set of Space-Division Multiple Access (SDMA) groups from each cluster. These groups are used as candidate groups to receive Scheduling Unit (SU) in the scheduling part. In order to compose a group, we employ a metric that takes into account the trade-off between the spatial channel correlation and channel gain of MSs. In this context, it is proposed a suboptimal solution to avoid the high complexity required by the optimal solution. Thirdly and finally, we use the candidate SDMA groups from the grouping part to solve the data rate maximization problem considering QoS requirements. The scheduling part can be solved by our proposed optimal solution based on Branch and Bound (BB). However, since it has high computational complexity, we propose a suboptimal scheduling algorithm that presents a reduced complexity. In the simulation results, we evaluate the performance of both optimal and suboptimal solutions, as well as an adaptation of the Joint Satisfaction Maximization (JSM) scheduler to a massive MIMO scenario. Although the suboptimal solution presents a performance loss compared to the optimal one, it is more suitable for practical settings as it is able to obtain a good performance-complexity trade-off. Furthermore, we show that the choice of a suitable trade-off between the spatial channel correlation and channel gain improves the system performance. Finally, for a low number of available SDMA groups, the suboptimal solution presents near optimal outage and a throughput loss of only 10% in comparison to the high-complexity optimal solution while it outperforms the JSM solution in terms of outage and system throughput.

The success of 5G networks relies on well-designed scheduling strategies that can tackle the increasing complexity of emerging mobile networks. Consequently, scheduling solutions are becoming more complicated, requiring high computational power and communication with very low latency. in this context, this work studies how to enhance existing scheduling solutions considering cellular networks with multi-connectivity, where users can have two or more simultaneous connections to different cells. Starting with a literature overview, we identify the main drawbacks of existing scheduling solutions for multi-connectivity scenarios. Then, we propose a novel scheduling concept that simplifies the scheduling task and overcomes several issues of existing scheduling solutions. The proposed solution logically splits the existing scheduling responsibilities across two simpler entities, thus it is named SRS. Furthermore, we propose a scheduling algorithm for practical SRS deployments, which is suitable for current and future multi-connectivity scenarios. Simulations show that the proposed solution has near-optimal performance in 5G multi-connectivity scenarios, when running in centralized and distributed architectures, while outperforming existing solutions by increasing the minimum user rate by 95 percent and the percentage of satisfied users by 50 percent at high system loads.

While the problems of sum-rate maximization and sum-power minimization subject to quality of service (QoS) constraints in the multiple input multiple output interference broadcast channel (MIMO IBC) have been widely studied, most of the proposed solutions have neglected the user scheduling aspect assuming that a feasible set of users has been previously selected. However, ensuring QoS for each user in the MIMO IBC involves the joint optimization of transmit/receive beamforming vectors, transmit powers, and user scheduling variables. To address the full problem, we propose a novel formulation of a rate-constrained sum-utility maximization problem which allows to either deactivate users or minimize the QoS degradation for some scheduled users in infeasible scenarios. Remarkably, this is achieved avoiding the complexity of traditional combinatorial formulations, but rather by introducing a novel expression of the QoS constraints that allows to solve the problem in a continuous domain. We propose centralized and decentralized solutions, where the decentralized solutions focus on practical design and low signaling overhead. The proposed solutions are then compared with benchmarking algorithms, where we show the effectiveness of the joint scheduling and transceiver design as well as the flexibility of the proposed solution performing advantageously in several MIMO IBC scenarios.

One of the most important functionalities for the dual connectivity (DC) technology in non-stand-alone (NSA) fifth generation (5G) networks is the flow control algorithm (FCA), which has the role of correctly splitting the data traffic of users in DC. However, delays in backhaul links and the time between two executions of the FCA may limit the effectiveness of the split ratios computed by the FCA in situations where unexpected channel variations or an increase in the traffic load happen. In this context, we propose a heuristic mechanism where the user assists the FCA in performing a faster and finer adjustment of the bearer split ratios. The proposed mechanism works based on a single scalar value and can be combined with any existing FCA. Besides the adjustment of the bearer split ratios, the proposed mechanism may also trigger handover/disconnection when it detects problems in one of the user connections, which is a novel feature that is not present in existing solutions. Simulations demonstrate the gains provided by the proposed mechanism in terms of user satisfaction, total system rate and user throughput.

In this article, we study a Radio Resource Allocation (RRA) that was formulated as a non-convex optimization problem whose main aim is to maximize the spectral efficiency subject to satisfaction guarantees in multiservice wireless systems. This problem has already been previously investigated in the literature and efficient heuristics have been proposed. However, in order to assess the performance of Machine Learning (ML) algorithms when solving optimization problems in the context of RRA, we revisit that problem and propose a solution based on a Reinforcement Learning (RL) framework. Specifically, a distributed optimization method based on multi-agent deep RL is developed, where each agent makes its decisions to find a policy by interacting with the local environment, until reaching convergence. Thus, this article focuses on an application of RL and our main proposal consists in a new deep RL based approach to jointly deal with RRA, satisfaction guarantees and Quality of Service (QoS) constraints in multiservice celular networks. Lastly, through computational simulations we compare the state-of-art solutions of the literature with our proposal and we show a near optimal performance of the latter in terms of throughput and outage rate.