IEEE Transactions on Network Science and Engineering最新文献

{"title":"Dataset Distillation-Based Hybrid Federated Learning on Non-IID Data","authors":"Xiufang Shi;Wei Zhang;Yuheng Li;Mincheng Wu;Zhenyu Wen;Shibo He;Tejal Shah;Rajiv Ranjan","doi":"10.1109/TNSE.2026.3679013","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3679013","url":null,"abstract":"In federated learning, the heterogeneity of client data has a great impact on the performance of model training. Many heterogeneity issues in this process are raised by non-independently and identically distributed (non-IID) data. To address the issue of label distribution skew, we propose a hybrid federated learning framework called HFLDD, which integrates dataset distillation to generate approximately independent and equally distributed (IID) data, thereby improving the performance of model training. In particular, we partition the clients into heterogeneous clusters, where the data labels among different clients within a cluster are unbalanced while the data labels among different clusters are balanced. The cluster heads collect distilled data from the corresponding cluster members, and conduct model training in collaboration with the server. This training process is like traditional federated learning on IID data, and hence effectively alleviates the impact of non-IID data on model training. We perform a comprehensive analysis of the convergence behavior, communication overhead, and computational complexity of the proposed HFLDD. Extensive experimental results based on multiple public datasets demonstrate that when data labels are severely imbalanced, the proposed HFLDD outperforms the baseline methods in terms of both test accuracy and communication cost.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8331-8347"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147696554","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Joint Optimization of Computation Offloading and Resource Allocation in Heterogeneous UAV-Assisted Edge Computing: A Game-Theoretical Approach","authors":"Jiale Wu;Xiaolong Xu;Guangming Cui;Jielin Jiang","doi":"10.1109/TNSE.2026.3680337","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3680337","url":null,"abstract":"In UAV-assisted edge computing systems, existing research inadequately addresses multi-resource heterogeneity, encompassing computational resource differences and energy consumption coefficient variations among edge servers. This heterogeneity causes critical load imbalance: energy-efficient servers attract excessive users despite limited capacity, while high-capacity servers with higher energy costs remain underutilized, degrading performance for latency-sensitive applications. To address this challenge, this paper formulates the joint offloading and resource allocation as an NP-hard optimization minimizing user costs, reducing resource competition, and maximizing server utility. We propose SOTNRA (Server Offloading Target Node selection and Resource Allocation), a hierarchical framework based on a two-layer game structure combining potential and Stackelberg games. The first stage employs a potential game-based server selection algorithm (PGSSA) considering load balancing constraints with proven convergence. The second stage uses a momentum-based gradient ascent algorithm (MBGA) for optimal pricing and offloading ratios. Extensive experiments demonstrate superior performance in reducing user costs while maintaining load balancing compared to single-dimensional heterogeneity approaches.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8348-8361"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147696584","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Reinforcement Learning Approach to Virtual Network Embedding Problems in 5G Networks","authors":"Amir Javadpour;Forough Ja'fari;Tarik Taleb;Chafika Benzaïd;Pedro R. Tomas;Luis Rosa;Jorge Proença;Luis Cordeiro","doi":"10.1109/TNSE.2026.3675357","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3675357","url":null,"abstract":"5G network slicing is the problem of mapping requested virtual networks on the substrate network resources. Due to resource capacity constraints, the performance of network slicing depends on the number of supported requests.This challenge is a type of Virtual Network Embedding (VNE) problem in which a weighted graph is divided into multiple smaller weighted graphs according to the user's custom requirements.These problems are NP-hard, and most existing solutions have suggested using Reinforcement Learning (RL) models to solve them. However, they do not adequately represent the weighted graph to the learning model. Therefore, their learning rate is limited. This paper proposes TRL-VNE, a Two-stage RL-based VNE solution to overcome these challenges. In the first stage of this solution, an RL model is utilized for mapping the central node of each request. Novel graph-based features (G-features) are used in this model to improve its learning rate. The second stage uses a greedy algorithm to map the other components. The simulation results show that TRL-VNE improves the requests acceptance ratio and maximum supported requests by 21% and 36%, respectively, compared to existing solutions. Moreover, we have proposed a network architecture based on TRL-VNE, and emulated it in Mininet to investigate the feasibility of the proposed solution.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8200-8223"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147696598","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Toward Robust PBFT Consensus: SMP Modeling and Grouping Optimization of Hierarchical Node Architecture","authors":"Junchao Fan;Yueqi Jiang;Bocheng Ju;Xiaolin Chang","doi":"10.1109/TNSE.2026.3678924","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3678924","url":null,"abstract":"Consensus mechanisms are essential to distributed networked systems, ensuring consistent state agreement under unreliable communication and node failures. Among such mechanisms, Practical Byzantine Fault Tolerance (PBFT) has been widely adopted in practice and extensively extended through numerous variants to address scalability in large-scale and heterogeneous networks. However, they still lack 1) adaptability to dynamic changes in network topology or fault distribution, and/or 2) capability of mitigating the gap between the consensus system behavior quantification and the consensus optimization approaches. These degrade consensus robustness in terms of consensus probability and consensus response time. To address these challenges, we first propose a hierarchical node architecture for PBFT. We then develop a scalable Semi-Markov Process (SMP)-based analytical model that captures the stochastic, time-dependent evolution of consensus states across multiple layers. Numerical and simulation results confirm the model's accuracy and show that the hierarchical structure sustains reliable operation up to 30% faulty nodes while maintaining low latency. Building on this modeling foundation, we propose a reinforcement learning-based optimization approach that formulates hierarchical grouping as a Markov decision process and solves it using proximal policy optimization. Experimental evaluation demonstrates that our approach improves consensus probability and reduces latency compared with static grouping baselines.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8261-8276"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147696654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multimessage Paxos With Priorities and Aging for Blockchain Applications","authors":"Elham Amini;Jelena Mišić;Vojislav B. Mišić","doi":"10.1109/TNSE.2026.3674834","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3674834","url":null,"abstract":"This paper extends the Paxos consensus algorithm by introducing preemptive priority scheduling with support for multiple proposal priorities, alongside an aging-based promotion mechanism to restore fairness and prevent starvation of lower-priority proposals. These enhancements address scenarios where proposals have heterogeneous urgency levels, which are common in blockchain and real-time distributed systems. Introducing priorities allows the protocol to expedite time-sensitive requests, while aging ensures that long-waiting low-priority proposals are eventually processed, thus improving fairness without compromising safety or liveness guarantees. We introduce a novel preemption-based aging mechanism that improves fairness based on actual contention dynamics rather than time. We analytically model the operation of the proposed algorithm using preemptive priority queueing disciplines, particularly the preemptive repeat different model, allowing for flexible and realistic handling of urgent proposals with variable service times. Our results demonstrate that multimessage handling combined with aging significantly improves both fairness and efficiency in consensus operation, especially under high-load conditions.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"7718-7736"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147606251","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Privacy-Aware Blockchain Approach for Efficient and Secure E-Health Applications","authors":"Hussien AbdelRaouf;Mahmoud Abouyoussef;Mostafa Fouda;Zubair Md Fadlullah;Mohamed I. Ibrahem;Rongxing Lu","doi":"10.1109/TNSE.2026.3683944","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3683944","url":null,"abstract":"The integration of the Internet of Things (IoT) in E-healthcare has enabled heart attack detection (HAD) leveraging artificial intelligence (AI). However, existing approaches face computational complexity, limited accuracy, and a lack of secure bidirectional communication between patients and healthcare providers (<inline-formula><tex-math>$mathcal {HP}$</tex-math></inline-formula>s) while preserving patients' privacy. While homomorphic encryption and secure multi-party computation can protect privacy, they incur substantial overhead. To overcome these limitations, {we propose a lightweight HAD methodology that ensures the preservation of security and privacy while maintaining performance reliability.} {First, we introduce a customized consortium blockchain harnessing group signatures to ensure patient anonymity and data unlinkability, enabling patients to securely communicate with <inline-formula><tex-math>$mathcal {HP}$</tex-math></inline-formula>s through unique and untraceable identifiers.} Next, we develop a lightweight HAD model via knowledge distillation (KD) from a complex high-performing model that captures spatialtemporal patterns and emphasizes key features in health data. Then, we devised a functional encryption (FE)-based cryptosystem to protect patients' data during model execution. Lastly, we advance explainable AI to highlight influential health features and enhance clinical trust. Evaluated on a real-world heart dataset and deployed on a testbed, our HAD model achieves 99.22% accuracy and a 99.23% F1-score, outperforming state-of-the-art techniques while reducing memory footprint, parameter count, and inference time by {98.19%}, {98.58%}, and {60.18%}, respectively, and fostering interpretability and trust in clinical decision-making. {Furthermore, the proposed blockchain with the FE approach enables secure model execution, protects patient data, and scales to 500,000 patients in under 2.5 minutes, while minimizing communication and computational overhead, achieving {95.43%} and {89.06%} reductions, respectively.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8628-8645"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147796062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hybrid Orchestration of AI Services and Microservices in Multi-Edge Collaboration Based on Lagrangian Relaxation Deep Reinforcement Learning","authors":"Jiaxiang Xu;Zhaoyi Wang;Di Han;Yi Hu;Fuming Fan;Menglan Hu;Kai Peng","doi":"10.1109/TNSE.2026.3681584","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3681584","url":null,"abstract":"The rapid development of artificial intelligence has promoted the application of AI services in various fields. The combination of the microservice architecture and the “AI as a Service” (AaaS) to form an intelligent microservice system can realize the orchestration of intelligent applications. In intelligent applications, AI services often require the key functions of accompanying microservices, such as service discovery, data collection, gateway routing, etc., to ensure the stability and efficiency. Otherwise, it will lead to the incompleteness of applications, severely affecting the overall performance of the system. Therefore, fine-grained hybrid orchestration of AI services and microservices is crucial. Nevertheless, it faces major challenges in the resource-constrained edge environment: complex dependencies, frequent data communication, and heterogeneous resource competition between microservices and AI services. Moreover, the tight coupling between deployment and routing leads to complex joint optimization problems. Considering either deployment or routing separately will lead to poor system performance. However, most of the existing studies ignore the above problems. To this end, this paper investigates the hybrid orchestration problem. First, we model intelligent applications as hybrid service invocation chains. Then, based on Jackson queuing networks and multi-instance models, we deeply study the complex dependencies among hybrid services and the heterogeneous resource allocation problem, and conduct an accurate analysis of the request delay and system energy consumption. Furthermore, we innovatively propose Lagrangian Relaxed Deep Deterministic Policy Gradient (LR_DDPG), which introduces a cost mechanism and transforms the constrained hybrid orchestration process into an unconstrained optimization problem by Lagrange relaxation to find the optimal deployment and routing strategy under complex constraints. Finally, the experimental results show that our algorithm achieves significant advantages in terms of latency, energy consumption, and request success rate.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"8820-8837"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147828827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Low-Rank Tensors and 3DTV Minimization Based Spectrum Cartography","authors":"Xin Wang;Bin Shen;Xiaoge Huang;Qianbin Chen","doi":"10.1109/TNSE.2026.3669372","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3669372","url":null,"abstract":"Tensor completion-based spectrum cartography (SC) in wireless networks has been extensively studied in recent years. Nevertheless, most existing algorithms are confined to two-dimensional (2D) geographical regions and entirely overlook the inherent local spatial smoothness (LSS) of the spectrum tensors. As the demand for three-dimensional (3D) SC continues to burgeon, direct adaptation of 2D region-based tensor completion algorithms to 3D scenarios either leads to inherent accuracy degradation or poses significant technical hurdles in precisely processing 3D-structured spatial data. To address these formidable challenges, a novel 3D spatial SC method is proposed in this paper. Firstly, the spectrum data is modeled as a fourth-order spectrum tensor, comprehensively capturing spatial information across the 3D space and multiple frequency bands. Subsequently, by leveraging the low-rank property of spectrum tensors and incorporating the 3DTV regularization as an essential component in the optimization problem, the proposed method effectively overcomes the critical limitation of conventional SC algorithms, which fail to account for LSS in tensor completion. Finally, two algorithms are proposed to solve the resulting optimization problem: 3DTV-based parallel matrix factorization (3DTV-PMF), and its variant with Cholesky-based inversion, named as 3DTV-PMFC. Simulation results demonstrate that the proposed algorithms achieve a minimum performance improvement of 16.15<inline-formula><tex-math>$%$</tex-math></inline-formula> compared to the conventional PMF algorithm.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"7465-7479"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sensing With Unknown Signals: ISAC Enabled Distributed Passive Sensing for Multi-Target Detection and Localization","authors":"Ruiqi Liu;Leyi Zhang;Yuanshuo Gang;Tianqi Mao;Qingqing Wu;Abbas Jamalipour","doi":"10.1109/TNSE.2026.3670825","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3670825","url":null,"abstract":"Integrated sensing and communication (ISAC) has been recognized as a critical usage scenario for the 6th generation (6G) networks, covering a vast variety of use cases. As wireless systems are densely deployed, signals transmitted by network nodes are pervasive in the environment, making it possible to opportunistically receive these signals and exploit them for passive sensing. This motivates distributed passive sensing systems that consist only of receivers and rely on signals emitted by noncooperative transmitters. In this paper, a distributed passive sensing system that operates with completely unknown signals is proposed and studied. As a first step, a blind channel identification based method is proposed for target detection, where the presence of targets is inferred from peaks in the estimated equivalent channel response. Then, the system functionality is extended to support multitarget localization. The resulting measurements across distributed receivers lead to a data association problem, for which we formulate a suitable model and propose a coping strategy to correctly associate observations. Numerical results demonstrate the feasibility and accuracy of the proposed method enabling target detection and localization using unknown signals and singleantenna, lowcost receivers.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"7599-7613"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147557998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Constraint-Driven Evolution of Multimodal Video Intelligence: A Network and System Perspective","authors":"Xuzhao Li;Xuchen Li;Shiyu Hu;Zhaorui Zhang;Kang Hao Cheong","doi":"10.1109/TNSE.2026.3668404","DOIUrl":"https://doi.org/10.1109/TNSE.2026.3668404","url":null,"abstract":"Multimodal video intelligence is increasingly shaped not only by algorithmic progress but also by the constraints of networked deployment, such as limited compute and communication resources, real-time latency, energy, and cost. This survey examines the field from a network and system perspective and organizes recent advances into a three-stage, constraint-driven trajectory that links algorithmic capabilities with deployment architectures and network behaviors. The first stage, Perception Enhancement, focuses on real-time closed-loop processing through lightweight models and edge-side control. The second stage, Semantic Understanding, addresses long-sequence and multimodal analysis by leveraging edge–cloud collaboration and distributed scheduling. The third stage, Cognitive Reasoning, supports causal and interactive intelligence via retrieval augmentation, long-term memory, and multi-agent coordination. Across these stages, algorithm–system co-design is reflected in network behaviors such as caching, compression, representation sharing, and cross-modal alignment. Overall, this survey highlights a shift from algorithm-centric design toward network-aware intelligence, from static perception to open-world generalization, and from isolated inference to networked interactive systems, providing insights for bridging AI research with real-world networked applications.","PeriodicalId":54229,"journal":{"name":"IEEE Transactions on Network Science and Engineering","volume":"13 ","pages":"7667-7683"},"PeriodicalIF":7.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147606208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}