A reinforcement learning approach to power control and rate adaptation in cellular networks

Ghadimi, Euhanna; Calabrese, Francesco; Peters, Gunnar; Soldati, Pablo

doi:10.1109/icc.2017.7997440

Cited by 94 publications

(62 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The power allocation problem has also been studied in [82], [83], which consider a cellular network with several cells and the base station in each cell serves multiple users. All base stations share the same frequency spectrum, which is further divided into orthogonal sub-bands within the cell for each user, i.e., there exists inter-cell interference but no intracell interference.…”

Section: B Power Allocation In Wireless Networkmentioning

confidence: 99%

“…We have iterated throughout the article that such simultaneous actions of all learning agents tend to make the environment observed by each agent highly nonstationary and compromises stability of DQN training. To address the issue, either the experience replay technique that is central to the success of deep RL is disabled as in [74], or the agents take turns to update their actions as in [22], [82], [85].…”

Section: Joint Spectrum and Power Allocation: Application Example mentioning

confidence: 99%

See 1 more Smart Citation

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

et al. 2020

View full text Add to dashboard Cite

It has been a long-held belief that judicious resource allocation is critical to mitigating interference, improving network efficiency, and ultimately optimizing wireless communication performance. The traditional wisdom is to explicitly formulate resource allocation as an optimization problem and then exploit mathematical programming to solve the problem to a certain level of optimality. Nonetheless, as wireless networks become increasingly diverse and complex, e.g., in the highmobility vehicular networks, the current design methodologies face significant challenges and thus call for rethinking of the traditional design philosophy. Meanwhile, deep learning, with many success stories in various disciplines, represents a promising alternative due to its remarkable power to leverage data for problem solving. In this paper, we discuss the key motivations and roadblocks of using deep learning for wireless resource allocation with application to vehicular networks. We review major recent studies that mobilize the deep learning philosophy in wireless resource allocation and achieve impressive results. We first discuss deep learning assisted optimization for resource allocation. We then highlight the deep reinforcement learning approach to address resource allocation problems that are difficult to handle in the traditional optimization framework. We also identify some research directions that deserve further investigation.

show abstract

Section: B Power Allocation In Wireless Networkmentioning

confidence: 99%

Section: Joint Spectrum and Power Allocation: Application Example mentioning

confidence: 99%

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

et al. 2020

View full text Add to dashboard Cite

show abstract

“…This work does not assume an explicit knowledge of the state transition probabilities. Here, we leverage principles of RL to optimize the transmit beam in a totally decentralized manner [6], [12], [19].…”

Section: Problem Formulationmentioning

confidence: 99%

“…Recently, the problem of low latency communication [9] and URLLC [10], [11] for 5G mmWave network was studied to evaluate the performance under the impact of traffic dispersion and network densification. Moreover, a reinforcement learning (RL) approach to power control and rate adaptation was studied in [12]. All these works focus on maximizing the time average of network throughput or minimizing the mean delay without providing any guarantees for higher order moments (e.g., variance, skewness, kurtosis, etc.).…”

Section: Introductionmentioning

confidence: 99%

Ultra-Reliable Communication in 5G mmWave Networks: A Risk-Sensitive Approach

Bennis

Debbah

et al. 2018

IEEE Commun. Lett.

View full text Add to dashboard Cite

Abstract-In this letter, we investigate the problem of providing gigabit wireless access with reliable communication in 5G millimeter-Wave (mmWave) massive multiple-input multipleoutput (MIMO) networks. In contrast to the classical network design based on average metrics, we propose a distributed risk-sensitive reinforcement learning-based framework to jointly optimize the beamwidth and transmit power, while taking into account the sensitivity of mmWave links due to blockage. Numerical results show that our proposed algorithm achieves more than 9 Gbps of user throughput with a guaranteed probability of 90%, whereas the baselines guarantee less than 7.5 Gbps. More importantly, there exists a rate-reliability-network density tradeoff, in which as the user density increases from 16 to 96 per km 2 , the fraction of users that achieve 4 Gbps are reduced by 11.61% and 39.11% in the proposed and the baseline models, respectively.

show abstract

“…In [8], the authors apply support vector machine approaches to optimize the transmit antenna selection. In [9], the authors considered deep reinforcement learning in resource management. In [10], deep-learning based approaches were applied in cache-enabled heterogeneous networks.…”

Section: Introductionmentioning

confidence: 99%

Efficient Minimum-Energy Scheduling with Machine-Learning Based Predictions for Multiuser MISO Systems

Lei

You

et al. 2018

2018 IEEE International Conference on Communications (ICC)

View full text Add to dashboard Cite

We address an energy-efficient scheduling problem for practical multiple-input single-output (MISO) systems with stringent execution-time requirements. Optimal user-group scheduling is adopted to enable timely and energy-efficient data transmission, such that all the users' demand can be delivered within a limited time. The high computational complexity in optimal iterative algorithms limits their applications in real-time network operations. In this paper, we rethink the conventional optimization algorithms, and embed machine-learning based predictions in the optimization process, aiming at improving the computational efficiency and meeting the stringent executiontime limits in practice, while retaining competitive energy-saving performance for the MISO system. Numerical results demonstrate that the proposed method, i.e., optimization with machinelearning predictions (OMLP), is able to provide a time-efficient and high-quality solution for the considered scheduling problem. Towards online scheduling in real-time communications, OMLP is of high computational efficiency compared to conventional optimal iterative algorithms. OMLP guarantees the optimality as long as the machine-learning based predictions are accurate.

show abstract

A reinforcement learning approach to power control and rate adaptation in cellular networks

Cited by 94 publications

References 23 publications

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

Ultra-Reliable Communication in 5G mmWave Networks: A Risk-Sensitive Approach

Efficient Minimum-Energy Scheduling with Machine-Learning Based Predictions for Multiuser MISO Systems

Contact Info

Product

Resources

About