Optimal Wireless Resource Allocation With Random Edge Graph Neural Networks
Open accessThe DOI matches its Crossref record; 40 of the 41 references it cites were found in a registry or an index.
Abstract
We consider the problem of optimally allocating resources across a set of transmitters and receivers in a wireless network. The resulting optimization problem takes the form of constrained statistical learning, in which solutions can be found in a model-free manner by parameterizing the resource allocation policy. Convolutional neural networks architectures are an attractive option for parameterization, as their dimensionality is small and does not scale with network size. We introduce the random edge graph neural network (REGNN), which performs convolutions over random graphs formed by the fading interference patterns in the wireless network. The REGNN-based allocation policies are shown to retain an important permutation equivariance property that makes them amenable to transference to different networks. We further present an unsupervised model-free primal-dual learning algorithm to train the weights of the REGNN. Through numerical simulations, we demonstrate the strong performance REGNNs obtain relative to heuristic benchmarks and their transference capabilities.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Graph neural network (GNN) has recently been shown to be able to harness the benefits of generalizability, interpretability, robustness, scalability, superior performance, real-time and distributed implementation for learning to optimize nonconvex problems, including power control [270], beamforming [23], and phase shift design [25].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [270]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 2×
“In fast fading scenarios the instantaneous value of the reward r(t) in (1) is not directly experienced by end users.”
BackgroundBertsekas, D. P. (1997) · Journal of the Operational Research Society · doi:10.1038/sj.jors.2600425 · Cited by 11,726
(1999) · Cited by 5,147
Wide & Deep Learning for Recommender Systems
Cheng, H.-T., Koc, L., Harmsen, J., Shaked, T., Chandra, T., Aradhye, H., Anderson, G., Corrado, G., Chai, W., Ispir, M., et al. (2016) · Proceedings of the 1st Workshop on Deep Learning for Recommender Systems · doi:10.1145/2988450.2988454 · Cited by 3,535
Shi, Q., Razaviyayn, M., Luo, Z.-Q., & He, C. (2011) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2011.2147784 · Cited by 2,129
Dual methods for nonconvex spectrum optimization of multicarrier systems
Yu, W., & Lui, R. (2006) · IEEE Transactions on Communications · doi:10.1109/tcomm.2006.877962 · Cited by 1,572 · cited 1×
“Some simplification is attained by working in the Lagrangian dual domain [1], [2] and subsequently using dual descent methods [3]–[5], or, alternatively, with heuristic optimization and scheduling methods [6]–[9].”
MethodDeep Convolutional Networks on Graph-Structured Data
Henaff, M., Bruna, J., & LeCun, Y. (2015) · arXiv · doi:10.48550/arxiv.1506.05163 · Cited by 1,255
Learning to Optimize: Training Deep Neural Networks for Interference Management
Sun, H., Chen, X., Shi, Q., Hong, M., Fu, X., & Sidiropoulos, N. D. (2018) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2018.2866382 · Cited by 874
Shi, Q., Razaviyayn, M., Luo, Z.-Q., & He, C. (2011) · 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2011.5946304 · Cited by 873
Cited by
Items 201–300 of 344
Learning Hybrid Precoding Efficiently for mmWave Systems with Mathematical Properties
Liu, S., Guo, J., & Yang, C. (2022) · Globecom · doi:10.1109/globecom48099.2022.10000790 · Cited by 8
“For example, a variety of wireless policies are with permutation equivariance (PE) property, which can be harnessed by graph neural networks (GNNs), say for learning power control/allocation of interference networks [10]–[12] and learning precoding of intelligent reflecting surface systems [13].”
BackgroundGraph Neural Networks for Physical-Layer Security in Multi-User Flexible-Duplex Networks
Perera, T., Atapattu, S., Fang, Y., & Evans, J. (2024) · International Conference on Communications · doi:10.1109/icc51166.2024.10622731 · Cited by 5
“Recent GNN applications encompass decentralized inference [18], resource allocation with random edge GNNs [19], and link scheduling [20].”
BackgroundLearning State-Augmented Policies for Information Routing in Communication Networks
Das, S., Naderializadeh, N., & Ribeiro, A. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3516556 · Cited by 6
“Similar to [3], [4], [5], [7], [8], [9], [10], [11], [33], we consider the problem of network utility maximization subject to multiple constraints.”
InfluentialSelf-citationLearning Precoding in Multi-User Multi-Antenna Systems: Transformer or Graph Transformer?
Duan, Y., Guo, J., & Yang, C. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3624290 · Cited by 8
“…stem from harnessing the permutation properties (say one-dimensional (1D)- or two-dimensional (2D)- permutation equivariance (PE) property) that broadly exist in wireless policies, i.e., the mappings from known parameters (say channel matrix) to decisions (say precoding matrix) [9]– [11], [17].”
Yu, Y. (2021) · Mobile Information Systems · doi:10.1155/2021/1684726 · Cited by 5
IGNNITION: fast prototyping of graph neural networks for communication networks
Pujol-Perich, D., Suárez‐Varela, J., Ferriol-Galmés, M., Wu, B., Xiao, S., Cheng, X., Cabellos‐Aparicio, A., & Barlet‐Ros, P. (2021) · ACM SIGCOMM Poster and Demo Sessions · doi:10.1145/3472716.3472853 · Cited by 6
“, forwarding devices, links, paths), and relations between them sequentially arranged [5, 6, 9, 10].”
BackgroundMethodRobust MIMO Detection using Hypernetworks with Learned Regularizers
Zilberstein, N., Dick, C., Doost-Mohammady, R., Sabharwal, A., & Segarra, S. (2022) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco55093.2022.9909722 · Cited by 5
Optimized Resource Allocation and Queue Management for Traffic Control in MANET
Ambika, I., Bhatia, S., Basheer, S., & Dadheech, P. (2022) · Computer Systems Science and Engineering · doi:10.32604/csse.2023.030786 · Cited by 5
Wang, Z., & Gombolay, M. (2022) · Proceedings of the International Conference on Automated Planning and Scheduling · doi:10.1609/icaps.v32i1.19839 · Cited by 7
“Specifically, GNNs have shown success in addressing resource optimization and scheduling problems, such as scheduling data processing clusters (Mao et al. 2019), multi-robot systems (Wang and Gombolay 2020), wireless networks (Eisen and Ribeiro 2020), and more (Ma et al. 2020).”
BackgroundPrimal Dual PPO Learning Resource Allocation in Indoor IRS-Aided Networks
Zhang, H., Liu, X., Long, K., & Poor, H. V. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685203 · Cited by 5
Energy-Efficient Power Allocation in Wireless Networks using Graph Neural Networks
Li, B., Verma, G., Rao, C. R., & Segarra, S. (2021) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf53345.2021.9723417 · Cited by 5
Learning with Multigraph Convolutional Filters
Butler, L., Parada-Mayorga, A., & Ribeiro, A. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095028 · Cited by 8
“However, they are insufficient for the modeling of heterogeneous domains [3, 4].”
BackgroundMethodSelf-citationGraph Neural Networks Based Channel Estimation for mmWave Massive MIMO Systems
Ye, M., Liang, X., Pan, C., Xu, Y., Jiang, M., & Li, C. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3589434 · Cited by 4
“According to the 3GPP TR 38.901 Release 15 channel model [44], we generate 6000, 1000, and 1000 channel samples as the training set, testing set, and validation set, respectively.”
InfluentialGNN-Enabled Coordinated Beamforming Design for High Speed Railway Communication Systems
He, C., Lu, Y., Zhang, R., Li, Y., Ai, B., & Niyato, D. T. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3649999 · Cited by 4
“Meanwhile, [20] extended REGNN to optimize system utility in wireless ad-hoc or cellular networks.”
Bhairanatti, S., & Kumar, S. M. (2023) · International Journal of Electrical and Electronics Engineering · doi:10.14445/23488379/ijeee-v10i1p103 · Cited by 4
Graph reinforcement learning with relational priors for predictive power allocation
Zhao, J., & Yang, C. (2025) · Science China Information Sciences · doi:10.1007/s11432-024-4261-1 · Cited by 5
“GNNs have been employed to optimize resource allocation in a variety of wireless systems [11–30].”
InfluentialXue, Z., & Liang, Z. (2025) · IISE Transactions · doi:10.1080/24725854.2025.2526782 · Cited by 4
“Eisen and Ribeiro (2020) The CNG underscores the limitations of traditional stochastic graphs and field theories in modeling complex spatial coverage relationships in wireless cellular networks, necessitating innovative approaches.”
Heterogeneous Resource Allocation With Multi-Task Learning for Wireless Networks
Mitsiou, N. A., Bouzinis, P. S., Sarigiannidis, P. G., & Karagiannidis, G. K. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3630058 · Cited by 4
“In one approach called the random edge GNN (REGNN) [23], convolutions were applied over random graphs representing fading interference patterns in wireless networks.”
Graphon Filters: Signal Processing in Very Large Graphs
Ruiz, L., Chamon, L. F. O., & Ribeiro, A. (2020) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco47968.2020.9287695 · Cited by 4
Self-citationStability Analysis of Unfolded WMMSE for Power Allocation
Chowdhury, A., Gama, F., & Segarra, S. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9747310 · Cited by 5
Zero-Shot Recurrent Graph Neural Networks for Beam Prediction in Non-Terrestrial Networks
Geng, Z., She, C., Zhang, D., Li, C., Li, Y., & Vucetic, B. S. (2022) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps56602.2022.10008701 · Cited by 4
“4 with the GNN in [14], and develop a RGNN algorithm.”
MethodState-Augmented Information Routing In Communication Systems With Graph Neural Networks
Das, S., Naderializadeh, N., & Ribeiro, A. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447843 · Cited by 4
“Similarly to [2, 3, 4, 6], we formulate the aforementioned routing problem as the maximization of a network utility function subject to a set of constraints.”
BackgroundMethodSelf-citationGowtham, V., Schreiner, F., Corici, M., Magedanz, T., Zacarias, I., Drummond, A. C., & Jukan, A. (2023) · IEEE Future Networks World Forum (FNWF) · doi:10.1109/fnwf58287.2023.10520590 · Cited by 5
Behmandpoor, P., Moonen, M., & Patrinos, P. (2024) · IEEE Transactions on Signal and Information Processing over Networks · doi:10.1109/tsipn.2024.3487421 · Cited by 4
Event-Based Temporal Graph Neural Network for Radio Resource Management
Marwani, M., & Kaddoum, G. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3654880 · Cited by 4
“To address this, Random Edge Graph Neural Networks (REGNNs) were introduced for improved scalability and adaptability, though they struggled in dynamic environments [22].”
Resource Allocation in Hybrid Radio-Optical IoT Networks Using GNN With Multi-Task Learning
Hamrouni, A., Pollin, S., & Sallouha, H. (2026) · IEEE Transactions on Machine Learning in Communications and Networking · doi:10.1109/tmlcn.2026.3674017 · Cited by 3
“Eisen et al. [45] addressed the resource allocation in wireless networks through a model-free approach using Random Edge Graph Neural Networks (REG-NNs) that operate over interference-induced random graphs.”
Learning to Optimize QoS-Constrained Beamforming in Multi-User Systems: A Penalty-Dual Framework
Li, Y., Liu, Y., Xu, F., Shi, Q., & Chang, T. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3437668 · Cited by 5
DBSCAN Clustering for User Pairing in Wireless Networks
Mourya, S., Reddy, P., Amuru, S., & Kuchi, K. (2025) · International Conference on Communication Systems & Networks (COMSNETS) · doi:10.1109/comsnets63942.2025.10885692 · Cited by 3
“Recent advancements have explored the application of machine learning techniques to address such complex problems in wireless communication systems [2]–[7].”
A Comparative Analysis of Chest X-ray Examination with AI Enhancement Using XAI Techniques
Özkurt, C. (2025) · Balkan Journal of Electrical and Computer Engineering · doi:10.17694/bajece.1448546 · Cited by 3
Self-Attention-Based Real-Time Signal Detector for Communication Systems With Unknown Channel Models
Chen, L., & Sun, L. (2021) · IEEE Communications Letters · doi:10.1109/lcomm.2021.3082708 · Cited by 4
Unsupervised Wireless Network Model‐Assisted Abnormal Warning Information in Government Management
Sun, Y. (2021) · Journal of Sensors · doi:10.1155/2021/1614055 · Cited by 3
“Adding a timeline attribute to a model can detect periodic behavior [5].”
BackgroundMitropoulou, K., Kokkinos, P. C., Soumplis, P., & Varvarigos, E. (2022) · International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP) · doi:10.1109/csndsp54353.2022.9908022 · Cited by 3
“The use of graph-based ML techniques towards automated resource allocation is mostly unexplored, with a few notable exceptions that leverage edge graph neural networks for the construction of allocation policies [21] and the prediction of corresponding QoS performance indicators [22].”
MethodURLLC : Faster, Higher, Stronger, and Together
She, C., Duong, T. Q., Khosravirad, S. R., Popovski, P., Bennis, M., & Quek, T. Q. S. (2023) · Wiley eBooks · doi:10.1002/9781119818366.ch1 · Cited by 4
Decentralized Learning Framework for Hierarchical Wireless Networks: A Tree Neural Network Approach
Kim, M., Lee, H., Hwang, S., Kim, M., Debbah, M. A., & Lee, I. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3359674 · Cited by 3
“This property has triggered recent progresses in the GNN-based decentralized network management tasks [23], [24], [25], [34], [35], [36], [37].”
BackgroundInfluentialIoannou, I. I., Raspopoulos, M., Nagaradjane, P., Christophorou, C., Khalifeh, A. F., & Vassiliou, V. (2024) · Asian Conference on Communication and Networks (ASIANComNet) · doi:10.1109/asiancomnet63184.2024.10811034 · Cited by 3
GraphGPT: An Intent-Based Management System for Next Generation Networks
Gowtham, V., Schreiner, F., Rao, A., Shivaprasad, S., Corici, M., & Magedanz, T. (2025) · IEEE International Conference on Network Softwarization (NetSoft) · doi:10.1109/netsoft64993.2025.11080567 · Cited by 3
Throughput optimization in heterogeneous MIMO networks: a GNN-based approach
Wang, T.-Y., Zhou, H., Kannan, R., Swami, A., & Prasanna, V. K. (2022) · doi:10.1145/3565473.3569191 · Cited by 6
“We consider 16 types of links: SISO links, 𝑁 𝑡 × 1 MISO links, 1 × 𝑁 𝑟 SIMO links, and 𝑡 × 𝑁 𝑟 MIMO links, where 𝑁 𝑡 , 𝑁 𝑟 ∈ [ 2 , 4 , 8 ] .”
BackgroundGradient based Information Aggregation of GNN for Precoder Learning
Chen, S., Han, S., & Li, Y. (2023) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2023-fall60731.2023.10333802 · Cited by 6
Fusion of Domain Knowledge and Data-Driven Power Allocation Optimization Methods
Miao, C., Zhang, J., Huang, J., & Yang, J. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3549434 · Cited by 2
“Another category is the unsupervised learning-based approach [27], [28], [29], which uses a neural network to parameterize the power allocation function and sets the optimization objective as a loss function, which no longer relies on accurate training data and simplifies the computational process.”
Behmandpoor, P., Eisen, M., Patrinos, P., & Moonen, M. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3592774 · Cited by 2
“During policy training, these constraints are typically managed either by including a penalization term in the reward function [8], [28], [33], [34], or by using the primal-dual optimization method, which directly addresses constrained optimization problems [18], [19], [23], [24], [26].”
Self-citationKernel-Truncated Spline: Estimator Fleksibel untuk Regresi Nonparametrik
Hidayat, R. S. A., Muthahharah, I., & Resmawan, R. (2025) · Euler Jurnal Ilmiah Matematika Sains dan Teknologi · doi:10.37905/euler.v13i2.33062 · Cited by 2
Camargo, R. G., Wang, Z.-Y., & Ribeiro, A. (2025) · IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp55912.2026.11460522 · Cited by 3
“This natural correspondence makes GNNs a strong candidate for tackling complex tasks in wireless systems [1–4].”
InfluentialSelf-citationCost-Efficient Federated Learning in Massive IoT: A Physics-Inspired Graph Learning Approach
Zhao, L., Wu, D., He, K., Niu, H., & Zhou, L. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3634230 · Cited by 2
“Eisen and Ribeiro propose a random edge GNN for bandwidth allocation, whose message aggregation is conducted over graphs randomly generated by the radio interference pattern [21].”
Adaptive Resource Allocation for 6G Network Slicing via Hybrid CNN-LSTM Architecture
Batool, I., Fouda, M. M., Ibrahem, M. I., Ismail, M., & Fadlullah, Z. M. (2025) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2025-fall65116.2025.11310079 · Cited by 2
“…critical gaps persist in existing solutions: (1) Computational complexity barriers, with current Graph Neural Network (GNN) approaches [18], [28] and transformer methods [21], requiring prohibitive inference times ( 28 − 40 ms ) that exceed 6G requirements of < 1 ms for URLLC applications;…”
AI-Driven Power Optimization in 6G Networks: A Comprehensive Review
Mutar, H. A., & ALRubeei, I. R. N. (2026) · International Journal of Computational and Electronic Aspects in Engineering · doi:10.26706/ijceae.7.1.20260104 · Cited by 3
Fast State-Augmented Learning for Wireless Resource Allocation With Dual Variable Regression
Uslu, Y. B., Naderializadeh, N., Eisen, M., & Ribeiro, A. (2026) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2026.3717245 · Cited by 9
“Graph neural networks (GNNs) have been widely used in learning resource allocation policies for wireless networks owing to their permutation equivariance, transferability, and stability properties [9], [27], [34], [35].”
Self-citationGraph Neural Network Based Scheduling: Improved Throughput Under a Generalized Interference Model
Sambamoorthy, R., Mandalapu, J., Swamy, P. S., Jain, B., & Altman, E. (2021) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-030-92511-6_9 · Cited by 2
Learning Constrained Network Slicing Policies for Industrial Applications
Agostini, P., Tohidi, E., Kasparick, M., & Stańczak, S. (2022) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops53468.2022.9882147 · Cited by 2
“Building on the observation that the structure of resource allocation problems can be cast into statistical learning problems [9], machine learning techniques, such as reinforcement learning (RL), have been recently leveraged for learning model-free network slicing policies in a datadriven manner [10], [11].”
BackgroundResource Allocation in Heterogeneous Network with Supervised GNNs
Sun, Q., Zhang, Y., Wu, H., & Petrosian, O. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-36625-3_28 · Cited by 2
Peng, B., Besser, K.-L., Raghunath, R., & Jorswieck, E. A. (2023) · Globecom · doi:10.1109/globecom54140.2023.10437080 · Cited by 2
“…control [19] (however, it is not permutation-equivariant), a permutation-equivariant NN architecture has been applied to signal detection [20], and the permutation-2023 equivariant graphical neural network (GNN) has been applied to capacity maximization (different from our objective) [21]–[23].”
MethodGraph Neural Network Performance Analysis in the Cell-Free Massive MIMO Wireless Systems
Rahman, M. H., Sejan, M. A. S., Aziz, M. A., Tabassum, R., & Song, H. (2024) · International Joint Conference on Computer Science and Software Engineering · doi:10.1109/jcsse61278.2024.10613688 · Cited by 2
“In various applications, including resource management [10], [13], end-to-end communication [11], and MIMO detection [8], GNN-based techniques have shown promising results.”
BackgroundJoint Transmission for Industrial 5G-TSN Networks: A Deep Graph Clustering Approach
Li, M., Liu, H., Chen, C., Gu, F., Guo, S., & Guan, X. (2024) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/iccc62479.2024.10681834 · Cited by 2
“• Rand-S [15], where the scheduling policy (with the output of sampling rate and operating CG) is learned in a distributed manner by the policy gradient theorem with the individual and instantaneous cost average f i . age .”
MethodGu, Z., Park, J., & Choi, J. (2026) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2026.3679505 · Cited by 2
“Webrieflydiscussafewkeymethods.”
Codebook-Based Downlink Beamforming with Imperfect CSI Using Model-Driven Deep Learning
Schynol, L., Hemsing, M., & Pesavento, M. (2024) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf60004.2024.10942629 · Cited by 2
“Unlike QoS constraints based on expected values [32], our approach prevents a scenario where extreme cases dominate the distribution, ensuring that most users can achieve the target data rate log 2 (1 + γ i ) .”
Zhang, C., & Zhong, Y. (2025) · The International Arab Journal of Information Technology · doi:10.34028/iajit/22/4/4 · Cited by 2
Model-Based Deep Learning for QoS-Aware Rate-Splitting Multiple Access Wireless Systems
Zhang, H., Chen, M., Vahid, A., Ye, F., & Sun, H. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3601049 · Cited by 4
“Authors in [39] provided a solution based on Lagrangian of original function, and trained to learn dual variables to punish the violation constraints.”
Learning Power Allocation for Cellular Systems with Data Rate-based Deep Neural Network
Guo, J., & Yang, C. (2022) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc51071.2022.9771923 · Cited by 2
“For example, noticing that resource allocation policies are with permutation equivariance (PE) property, graph neural networks (GNNs) are designed in [9], [10], which either improve learning performance or reduce training complexity.”
BackgroundChen, H., Yang, G., Zhang, J., Yin, Q., & Huang, K. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892225 · Cited by 3
“Many important real-world applications naturally take the form of graphs, such as traffic flow [19], wireless communications [20], interfaces between proteins [21], and social networks [22].”
BackgroundGraph Neural Networks for Timely Updates of Short Packets in Interference-Limited Networks
Liu, Y., She, C., Hardjawana, W., & Vucetic, B. S. (2022) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf56349.2022.10051893 · Cited by 2
“Inspired by [8], we use the REGNN in the actor as an approximation of the power control policy, φ(·).”
MethodKumar, A., Shukla, R. K., & Shukla, R. S. (2023) · IEEE Delhi Section Conference (DELCON) · doi:10.1109/delcon57910.2023.10127306 · Cited by 2
“Therefore, in order to guarantee network connectivity and prolong network life, agile growth must be considered [11], [12].”
BackgroundAutoOPT: Data Generation and Optimization for Digital Twin Network (DTN)
Li, M., Zhou, C., Lu, L., Zhang, K., & Sun, T. (2023) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud60044.2023.00051 · Cited by 2
“AI models, especially neural networks, have been used to model complex network environments for scenarios such as performance evaluation [7] [8], resource allocation [9], and task scheduling [10], which helps build real-time, lightweight,”
MethodBehmandpoor, P., Patrinos, P., & Moonen, M. (2023) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco58844.2023.10290055 · Cited by 3
A Gradient Driven Graph Neural Network for Optimizing Precoding
Li, Z., Han, S., Yang, C., & Li, Y. (2023) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps58843.2023.10464444 · Cited by 2
“…was supported by the National Key R&D Program of China under Grant 2022YFB2902002 and the National Natural Science Foundation of China (NSFC) under Grant 62271024. property existed in many wireless policies [4,6], which have been designed to learn power allocation [6, 7] and precoding [5] policies.”
Reducing Sample Complexity of Deep Learning with Symmetric Prior of Wireless TasksPreprint
Sun, C., Wu, J., & Yang, C. (2020) · arXiv (Cornell University) · Cited by 2
Weighted Sum-Rate Maximization with Causal Inference for Latent Interference Estimation
You, L. (2023) · International Conference on Communications · doi:10.1109/icc45041.2023.10279554 · Cited by 2
“Unsupervised model training is also possible by in-cooperating learning in the processing of a primal-dual algorithm [10].”
MethodA Graph Neural Network Power Allocation Algorithm Based on Fully Unrolled WMMSE
Miao, C., Zhang, J., Huang, J., & Yang, J. (2024) · IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM) · doi:10.1109/wowmom60985.2024.00035 · Cited by 1
“The other category is based on unsupervised learning methods [9], [13], [14].”
BackgroundMethodLearning End-to-End Precoding for Time-Varying Channels with Graph Neural Networks
Liu, J., Liu, T., & Yang, C. (2024) · International Conference on Wireless Communications and Signal Processing (WCSP) · doi:10.1109/wcsp62071.2024.10827597 · Cited by 1
Gu, Z., Park, J., Vucetic, B. S., & Choi, J. D. (2026) · IEEE Transactions on Networking · doi:10.1109/ton.2026.3702752 · Cited by 1
“Further study is needed on relaxation-based slot assignment methods that minimize the number of slots while ensuring reliable transmissions in WTSN. Additionally, deep learning [27], [28] and spectral graph [29], [30] methods are also widely applied to solve graph problems.”
On the Transferability of Graph Neural Networks for Resource Allocation in Wireless Networks
Fernández, S., Camargo, R. G., Eisen, M., Ribeiro, A., & Larroca, F. (2024) · IEEE URUCON · doi:10.1109/urucon63440.2024.10850268 · Cited by 1
“Several other forms of the radio resource allocation problem exist, and it is rather straightforward to modify (1) to consider them (see for instance [7], [13]).”
Self-citationT³Planner: Multi-Phase Planning Across Structure-Constrained Optical, IP, and Routing Topologies
Hao, Y., Yang, S., Li, F., Zhang, Y., Zhao, C., Ren, X., Zhao, P. F., Xu, C., & Wang, S. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3543511 · Cited by 1
“GNNs are designed for structured data, and perform well in virtual network embedding [28], network modeling [29], and resource allocation [30].”
Gradient-Driven Graph Neural Networks for Learning Digital and Hybrid Precoder
Li, Z., Han, S., & Yang, C. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3628767 · Cited by 4
“Graph neural networks (GNNs) can leverage the permutation equivariant (PE) property existed in many wireless policies [9], [11], which has been exploited for learning tasks such as power allocation [11], [12] and precoding policies [10], [13].”
Peng, S., Liu, R., & Liu, Y. (2025) · International Conference on Machine Learning, Big Data and Business Intelligence (MLBDBI) · doi:10.1109/mlbdbi67855.2025.11331348 · Cited by 2
“In terms of specific applications, Eisen and Ribeiro [1] introduced a random-edge GNN framework that achieved near-optimal performance in wireless resource allocation while reducing computational complexity.”
Schynol, L., & Pesavento, M. (2026) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2026.3671023 · Cited by 1
“∀ S ∈ supp( P ( S ) ) Unlike QoS constraints based on expected values [55] or worst-case robust designs [15], outage constraints prevent extreme cases dominating the solution, thus ensuring that the majority of users achieve the target QoS level γ i .”
Resource Allocation in Fronthaul-Constrained Cell-Free Networks Using Edge-Graph Attention Networks
Liu, X., Zhao, J., Shen, F., Yang, K., & Sun, S. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3669944 · Cited by 2
“The authors of [22] proposed an unsupervised and model-free primal-duality learning algorithm to train REGNN, and proved that it can obtain superior performance and migration ability compared with the benchmark algorithms.”
GNN-Enhanced Binary Loop Detection for NOMA–AFDM
Li, Q., Liu, Y., Liu, G., Tian, D., Xu, F., & Liu, C. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3699661 · Cited by 1
“This is because the trainable weights within the GNN are shared across local node neighborhoods and are independent of the global graph dimension [49], [50].”
Xue, L., Dong, X., Hu, X., Zhang, Y., Yang, L., & Xiao, G. (2026) · Electronics · doi:10.3390/electronics15163661 · Cited by 1
Kalogerias, D. S., Eisen, M., Pappas, G. J., & Ribeiro, A. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053942 · Cited by 2
“As recently discussed in [9, 10], a particularly effective way to facilitate policy search in the stochastic variational program (1) is to restrict the class of admissible policies as where the constraint θ ∈ Θ ⊆ θ ∈ R N φ (cid:12)(cid:12) φ ( · , θ ) ∈ P (cid:9) is additionally enforced, for some…”
BackgroundSelf-citationSong, B., Zhou, Z., Li, C., Guo, D., Fu, X., & Hong, M. (2023) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc53906.2023.10304444 · Cited by 1
“For example, for multi-user power control problem, [9] uses feed forward networks (FFNs) to replace the WMMSE algorithm, [10] use graph neural networks (GNNs), while [11] adopts convolutional neural networks (CNNs).”
MethodDecentralized Provisioning Algorithm For BLE Mesh Network
Zhuravinskyi, M., MALETSKYI, D., Кравец, И. Б., Kapshii, O., & Farenyuk, O. (2023) · IEEE International Conference on Computer Sciences and Information Technologies (CSIT) · doi:10.1109/csit61576.2023.10324056 · Cited by 1
“Other similar algorithms were investigated and shown to have advantages, at least in some configurations [6], [7], [8].”
BackgroundA Study on Adversarial Machine Learning in Wireless Communication Systems
Ajayi, O. T., Onidare, S. O., & Tajudeen, H. (2024) · Lecture notes in electrical engineering · doi:10.1007/978-981-97-6937-7_46 · Cited by 1
FGNN-Based Improved Resource Distribution Framework for V2X Wireless Networks
Waqas, S. M., Tang, Y., Abbas, F., Hussain, M., & Bangash, Y. A. (2024) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2024-spring62846.2024.10683058 · Cited by 1
“Notably, recent studies, including [2], [7], [15], have presented a persistent formulation of the resource distribution challenge, enabling the incorporation of stochastic constraints in the optimization process, for instance, assigning the least capacity to every vehicle.”
BackgroundRomero-Barrera, A. J., Cruz-Piris, L., Tejedor‐Romero, M., Giménez-Guzmán, J. M., & Marsá-Maestre, I. (2024) · IEEE Access · doi:10.1109/access.2024.3523464 · Cited by 2
Graph Neural Network for Multi-User MISO Secure Wireless Communications
Zhao, K., Tang, X., Dong, L., Zhang, R., & Du, Q. (2025) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc61545.2025.10978207 · Cited by 1
“Besides, another feature of GNN is that GNN has permutation equivariant property [17], which means the index labels of nodes changing dosen’t affect the output of the neural network.”
Precoder Learning by Leveraging Unitary Equivariance Property
Ge, Y., Liao, S., Han, S., & Yang, C. (2025) · Globecom · doi:10.1109/globecom59602.2025.11432657 · Cited by 1
“Incorporating inductive biases into DNN architecture based on mathematical properties of the policies being learned can restrict the hypothesis space of DNNs and thus improve learning efficiency [5].”
Enhanced mobility management mechanisms for 5G networks
Jain, A. (2020) · UPCommons institutional repository (Universitat Politècnica de Catalunya) · doi:10.5821/dissertation-2117-330736 · Cited by 1
“Lastly, given the current research trends in Artificial Intelligence and Machine Learning, as well as certain important works in the area of MM utilizing the same [226], we believe it will be prudent to explore these techniques for the future MM strategies as well.”
BackgroundAmini, M. H. (2024) · The Research Repository @ WVU (West Virginia University) · doi:10.33915/etd.12402 · Cited by 3
Almost-Zero Duality Gaps in Model-Free Resource Allocation for Wireless Systems
Kalogerias, D. S., Eisen, M., Pappas, G. J., & Ribeiro, A. (2020) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco47968.2020.9287660
“However, the recent advent of machine learning for wireless communications [6]–[10] has dawned the realization that both these challenges can be ameliorated with the incorporation of learning parameterizations [6], [7].”
BackgroundSelf-citationDistributed Learning in Non-Convex World: Theory, Algorithms, and Applications
Sun, H. (2021) · University of Minnesota Digital Conservancy (University of Minnesota)
Cognitive Communications System for Ultra-Low Size, Weight and Power (SWAP) Attributable Platforms
LaMar, S., Fitting, R., & Jayasumana, A. P. (2022) · IEEE Access · doi:10.1109/access.2022.3167039
“Researchers have used these networks to solve n-body collision problems, networking problems to learn routing protocols [18], predict jitter and delay [19], optimize resource allocations [20], and perform distributed transmission scheduling.”
Graph-based Deep Learning for Spatial Reuse Optimization in Dense WLAN DeploymentsPreprint
Azeez, S., & Henna, S. (2022) · doi:10.36227/techrxiv.19391477.v1
Graph Based Deep Learning for Spatial Reuse Optimization in Dense WLAN DeploymentsPreprint
Azeez, S., & Henna, S. (2022) · doi:10.36227/techrxiv.19391477
Graph Based Deep Learning for Spatial Reuse Optimization in Dense WLAN DeploymentsPreprint
Azeez, S., & Henna, S. (2022) · doi:10.36227/techrxiv.19391477.v2
Graph Based Deep Learning for Spatial Reuse Optimization in Dense WLAN DeploymentsPreprint
Azeez, S., & Henna, S. (2022) · doi:10.36227/techrxiv.19391477.v3
Graph Convolutional Network Enabled Power-Constrained HARQ Strategy for URLLC
Chen, Y., Shi, Z., Wangy, H., Fuz, Y., Yang, G., Max, S., & Ding, H. (2023) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/icccworkshops57813.2023.10233739
“Instead of optimizing p, the neural network parameters W need to be optimally determined through the primal-dual learning approach [17].”
MethodMCSO-MW: An Energy Optimized Data Switching Scheme for Multi-hop Wireless Sensor Networks
Alagarsamy, A., Mahilmaran, S., Sureshkumar, K., & Vimala, P. (2023) · Journal of Physics Conference Series · doi:10.1088/1742-6596/2471/1/012008
“Works [13]-[15] explain the single-hop data communication, where the single hop between sensor nodes and TP leads to a critical routing path for PS.”
BackgroundResultSureshkumar, K., & Vimala, P. (2023) · Journal of Physics Conference Series · doi:10.1088/1742-6596/2471/1/012013
“The discussion on large-dimension networks [17], with fading and interference characteristics, states that transference is achieved with a proper permutation of independent parameters selected to achieve scalability.”
BackgroundOptimal Collaborative Uploading in Crowdsensing with Graph Learning
Zhang, Y., Luan, T. H., Wang, H., Wang, L., Yu, Z., Guo, B., & Zhao, Y. (2023) · International Conference on Communications · doi:10.1109/icc45041.2023.10279250
“Specifically, GNN-A2C means that we replace the NN model in our A2C framework with GNN module from [20].”
MethodGraph-based Selection-Activation Reinforcement Learning for Heterogenous Multi-agent Collaboration
Chen, H.-X., Zhang, X., & Shen, J.-N. (2024) · Chinese Control Conference (CCC) · doi:10.23919/ccc63176.2024.10662338
“With the ability to rapidly cluster valid information, Graph neural network (GNN) is now widely applied in various domains, including natural language processing [8], wireless communication [9], and data processing [10].”
BackgroundImproved Stability Bounds for Graph Convolutional Neural Networks Under Graph Perturbations
Zhang, J., & Zhao, Z. (2024) · IEEE Information Theory Workshop (ITW) · doi:10.1109/itw61385.2024.10806975
“To capture the underlying topology of data points in non-Euclidean spaces, graphs have been widely used across various applications, including social sciences [4], financial networks [5], wireless communications [6], molecular structures [7], among others.”
Learning to optimize via graph neural networks
Yong, Z., Zou, Y., Wu, Y., Shi, Y., & Zhang, J. (2025) · Elsevier eBooks · doi:10.1016/b978-0-44-322073-9.00020-1
Metadata
- Authors
- Eisen, M., & Ribeiro, A.
- Affiliations
- Intel (United States) (US)
- University of Pennsylvania (US)
- Venue
- IEEE Transactions on Signal Processing
- Published
- 2020
- Volume
- 68
- Pages
- 2977-2991
- ISSN
- 1053-587X (ISSN-L)
- 1941-0476
- OpenAlex
- W2971598376
- Cited by
- 327
- References
- 41
- Citations by year
- 2026: 31 · 2025: 53 · 2024: 59 · 2023: 60 · 2022: 56 · 2021: 44 · 2020: 23
- FWCI
- 15.72
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Signal Processing (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- ARL DCIST CRA (W911NF-17-2-0181)
- Intel Science and Technology Center for Wireless Autonomous Systems
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Wireless Network Optimization
- Physical Sciences › Computer Science › Computer Networks and Communications › Wireless Communication Networks Research
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Keywords
- wireless resource allocation, graph neural networks, permutation equivariance, model-free reinforcement learning, random graphs, convolutional neural networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.