Journal of Network and Computer Applications最新文献_第10页

PRAETOR:Packet flow graph and dynamic spatio-temporal graph neural network-based flow table overflow attack detection method PRAETOR：基于数据包流图和动态时空图神经网络的流表溢出攻击检测方法

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-17 DOI: 10.1016/j.jnca.2025.104333

Kaixi Wang , Yunhe Cui , Guowei Shen , Chun Guo , Yi Chen , Qing Qian

The flow table overflow attack on SDN switches is considered to be a destructive attack in SDN. By exhausting the computing and storage resources of SDN switches, this attack severely disrupts the normal communication functions of SDN networks. Graph neural networks are now being employed to detect flow table overflow attacks in SDN. When a flow graph is constructed, flow features are commonly utilized as nodes to represent the characteristics of flow table overflow attacks. However, a graph solely relying on these nodes and attributes may not fully encompass all the nuances of the flow table overflow attack. Additionally, GNN model may be difficult in capturing the graph information between different flow graphs over time, thus decreasing the detection accuracy of packet flow graph. To address these issues, we introduce PRAETOR, a detection method for flow table overflow attacks that leverages a packet flow graph and a dynamic spatio-temporal graph neural network. More particularly, The PaFlo-Graph algorithm and the EGST model are introduced by PRAETOR. The PaFlo-Graph algorithm generates a packet flow graph for each flow. It utilizes packet information to construct the graph with more detail, thereby better reflecting the characteristics of flow table overflow attacks. The EGST model is a dynamic spatio-temporal graph convolutional network designed to detect flow table overflow attacks by analyzing packet flow graphs. Experiments were conducted under two network topologies, where we used tcpreplay to replay packets from the bigFlow dataset to simulate SDN network flow. We also employed sFlow to sample packet features. Based on the sampled data, two datasets were constructed, each containing 1,760 network flows. For each packet, eight key features were extracted to represent its characteristics. The evaluation metrics include TPR, TNR, accuracy, precision, recall, F1-score, confusion matrix, ROC curves, and PR curves. Experimental results show that the proposed PaFlo-Graph algorithm generates more detailed flow graphs compared to KNN and CRAM, resulting in an average improvement of 6.49% in accuracy and 8.7% in precision. Furthermore, the overall detection framework, PRAETOR, achieves detection accuracies of 99.66% and 99.44% on Topo1 and Topo2, respectively. The precision scores reach 99.32% and 99.72%, and the F1-scores are 99.57% and 100%, respectively, indicating superior detection performance compared to other methods.

针对SDN交换机的流表溢出攻击被认为是SDN网络中的一种破坏性攻击。该攻击通过耗尽SDN交换机的计算和存储资源，严重破坏SDN网络的正常通信功能。图神经网络目前被用于检测SDN中的流表溢出攻击。在构建流图时，通常使用流特征作为节点来表示流表溢出攻击的特征。然而，仅仅依赖于这些节点和属性的图可能无法完全包含流表溢出攻击的所有细微差别。此外，随着时间的推移，GNN模型可能难以捕获不同流图之间的图形信息，从而降低了包流图的检测精度。为了解决这些问题，我们引入了PRAETOR，这是一种利用数据包流图和动态时空图神经网络的流表溢出攻击检测方法。具体地说，PRAETOR介绍了PaFlo-Graph算法和EGST模型。PaFlo-Graph算法为每个流生成数据包流图。它利用报文信息构造更详细的图，从而更好地反映了流表溢出攻击的特点。EGST模型是一个动态的时空图卷积网络，旨在通过分析数据包流图来检测流表溢出攻击。实验在两种网络拓扑下进行，其中我们使用tcpreplay来重播来自bigFlow数据集的数据包来模拟SDN网络流。我们还使用sFlow对数据包特征进行采样。基于采样数据，构建了两个数据集，每个数据集包含1760个网络流。对于每个数据包，提取8个关键特征来表示其特征。评价指标包括TPR、TNR、正确率、精密度、召回率、f1评分、混淆矩阵、ROC曲线、PR曲线。实验结果表明，与KNN和CRAM算法相比，本文提出的PaFlo-Graph算法生成的流图更加详细，准确率平均提高6.49%，精度平均提高8.7%。此外，整体检测框架PRAETOR在Topo1和Topo2上的检测准确率分别达到99.66%和99.44%。精密度得分达到99.32%、99.72%，f1得分分别达到99.57%、100%，检测性能优于其他方法。

{"title":"PRAETOR:Packet flow graph and dynamic spatio-temporal graph neural network-based flow table overflow attack detection method","authors":"Kaixi Wang , Yunhe Cui , Guowei Shen , Chun Guo , Yi Chen , Qing Qian","doi":"10.1016/j.jnca.2025.104333","DOIUrl":"10.1016/j.jnca.2025.104333","url":null,"abstract":"<div><div>The flow table overflow attack on SDN switches is considered to be a destructive attack in SDN. By exhausting the computing and storage resources of SDN switches, this attack severely disrupts the normal communication functions of SDN networks. Graph neural networks are now being employed to detect flow table overflow attacks in SDN. When a flow graph is constructed, flow features are commonly utilized as nodes to represent the characteristics of flow table overflow attacks. However, a graph solely relying on these nodes and attributes may not fully encompass all the nuances of the flow table overflow attack. Additionally, GNN model may be difficult in capturing the graph information between different flow graphs over time, thus decreasing the detection accuracy of packet flow graph. To address these issues, we introduce PRAETOR, a detection method for flow table overflow attacks that leverages a packet flow graph and a dynamic spatio-temporal graph neural network. More particularly, The PaFlo-Graph algorithm and the EGST model are introduced by PRAETOR. The PaFlo-Graph algorithm generates a packet flow graph for each flow. It utilizes packet information to construct the graph with more detail, thereby better reflecting the characteristics of flow table overflow attacks. The EGST model is a dynamic spatio-temporal graph convolutional network designed to detect flow table overflow attacks by analyzing packet flow graphs. Experiments were conducted under two network topologies, where we used tcpreplay to replay packets from the bigFlow dataset to simulate SDN network flow. We also employed sFlow to sample packet features. Based on the sampled data, two datasets were constructed, each containing 1,760 network flows. For each packet, eight key features were extracted to represent its characteristics. The evaluation metrics include TPR, TNR, accuracy, precision, recall, F1-score, confusion matrix, ROC curves, and PR curves. Experimental results show that the proposed PaFlo-Graph algorithm generates more detailed flow graphs compared to KNN and CRAM, resulting in an average improvement of 6.49% in accuracy and 8.7% in precision. Furthermore, the overall detection framework, PRAETOR, achieves detection accuracies of 99.66% and 99.44% on Topo1 and Topo2, respectively. The precision scores reach 99.32% and 99.72%, and the F1-scores are 99.57% and 100%, respectively, indicating superior detection performance compared to other methods.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104333"},"PeriodicalIF":8.0,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145121251","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}

引用次数: 0

ST-MemA: Leveraging Swin Transformer and memory-enhanced LSTM for encrypted traffic classification ST-MemA：利用Swin Transformer和内存增强的LSTM进行加密流量分类

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-17 DOI: 10.1016/j.jnca.2025.104329

Zhiyuan Li , Yujie Jin

Traffic classification is essential for effective intrusion detection and network management. However, with the pervasive use of encryption technologies, traditional machine learning-based and deep learning-based methods often fall short in capturing the fine-grained details in encrypted traffic. To address these limitations, we propose a memory-enhanced LSTM model based on Swin Transformer for multi-class encrypted traffic classification. Our approach first reconstructs raw encrypted traffic by converting each flow into single-channel images. A hierarchical attention network, incorporating both byte-level and packet-level attention, then performs comprehensive feature extraction on these traffic images. The resulting feature maps are subsequently classified to identify traffic flow categories. By combining the long-term dependency capabilities of LSTM with the Swin Transformer’s strengths in feature extraction, our model effectively captures global features across diverse traffic types. Furthermore, we enhance LSTM with memory attention, enabling the model to focus on more fine-grained information. Experimental results on three public datasets—USTC-TFC2016, ISCX-VPN2016, and CIC-IoT2022 show that our model, ST-MemA, improves the classification accuracy to 99.43%, 98.96% and 98.21% and

F_{1}

-score to 0.9936, 0.9826 and 0.9746, respectively. The results also demonstrate that our proposed model outperforms current state-of-the-art models in classification accuracy and computational efficiency.

流分类是有效的入侵检测和网络管理的基础。然而，随着加密技术的广泛使用，传统的基于机器学习和深度学习的方法往往无法捕获加密流量中的细粒度细节。为了解决这些限制，我们提出了一种基于Swin Transformer的内存增强LSTM模型，用于多类加密流分类。我们的方法首先通过将每个流转换为单通道图像来重建原始加密流量。然后，结合字节级和包级注意的分层注意网络对这些流量图像进行全面的特征提取。随后对得到的特征图进行分类，以确定交通流类别。通过将LSTM的长期依赖能力与Swin Transformer在特征提取方面的优势相结合，我们的模型有效地捕获了不同流量类型的全局特征。此外，我们通过内存关注来增强LSTM，使模型能够关注更细粒度的信息。在ustc - tfc2016、ISCX-VPN2016和CIC-IoT2022三个公共数据集上的实验结果表明，我们的ST-MemA模型将分类准确率分别提高到99.43%、98.96%和98.21%，f1得分分别提高到0.9936、0.9826和0.9746。结果还表明，我们提出的模型在分类精度和计算效率方面优于当前最先进的模型。

{"title":"ST-MemA: Leveraging Swin Transformer and memory-enhanced LSTM for encrypted traffic classification","authors":"Zhiyuan Li , Yujie Jin","doi":"10.1016/j.jnca.2025.104329","DOIUrl":"10.1016/j.jnca.2025.104329","url":null,"abstract":"<div><div>Traffic classification is essential for effective intrusion detection and network management. However, with the pervasive use of encryption technologies, traditional machine learning-based and deep learning-based methods often fall short in capturing the fine-grained details in encrypted traffic. To address these limitations, we propose a memory-enhanced LSTM model based on Swin Transformer for multi-class encrypted traffic classification. Our approach first reconstructs raw encrypted traffic by converting each flow into single-channel images. A hierarchical attention network, incorporating both byte-level and packet-level attention, then performs comprehensive feature extraction on these traffic images. The resulting feature maps are subsequently classified to identify traffic flow categories. By combining the long-term dependency capabilities of LSTM with the Swin Transformer’s strengths in feature extraction, our model effectively captures global features across diverse traffic types. Furthermore, we enhance LSTM with memory attention, enabling the model to focus on more fine-grained information. Experimental results on three public datasets—USTC-TFC2016, ISCX-VPN2016, and CIC-IoT2022 show that our model, ST-MemA, improves the classification accuracy to 99.43%, 98.96% and 98.21% and <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span>-score to 0.9936, 0.9826 and 0.9746, respectively. The results also demonstrate that our proposed model outperforms current state-of-the-art models in classification accuracy and computational efficiency.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104329"},"PeriodicalIF":8.0,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145094148","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}

引用次数: 0

Fault-tolerant 3-D topology construction of UAV-BSs for full coverage of users with different QoS demands UAV-BSs的容错三维拓扑构建，实现对不同QoS需求用户的全覆盖

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-17 DOI: 10.1016/j.jnca.2025.104332

Xijian Luo , Jun Xie , Liqin Xiong , Yaqun Liu , Yuan He

Deploying Unmanned aerial vehicle mounted base stations (UAV-BSs) in post-disaster areas or battlefields, where the ground infrastructures are missing or destroyed, can quickly restore communication coverage. Due to the unstable and hostile properties of the environments, the ability to maintain the connectivity of the UAV-BSs network should be considered. In this paper, we study the deployment of UAV-BSs to provide full coverage for users with different quality of service (QoS) demands. The objective is to minimize the number of UAV-BSs under the constraints of user demands and UAV-BS service abilities. Besides, in absence of ground base stations, we also aim to construct a bi-connected topology for the UAV-BS network. However, the formulated problem, as a special instance of the geometric disk cover (GDC) problem, is NP-hard. To tackle this problem, we propose a heuristic algorithm, named Improved QoS-Prior Coverage and bi-Connectivity (IQP2C), by separately solving the user coverage and bi-connected topology construction subproblems. Firstly, IQP2C provides full coverage for users with minimum covering UAVs. Then, we propose an altitude-cluster-based method extending from the 2-D Hamilton cycle to construct bi-connectivity for the UAV-BS network. Simulation results validate the effectiveness of IQP2C in meeting different QoS demands and constructing fault-tolerant topology. Moreover, IQP2C outperforms other baselines in terms of minimized number of UAV-BSs for user coverage, minimized number of UAV-BSs for bi-connectivity as well as running time.

在地面基础设施缺失或被破坏的灾后地区或战场部署无人机基站（UAV-BSs），可以快速恢复通信覆盖。由于环境的不稳定和敌对性质，应考虑保持UAV-BSs网络连通性的能力。在本文中，我们研究了部署UAV-BSs以实现对不同服务质量（QoS）需求的用户的全覆盖。目标是在用户需求和无人机- bs服务能力的约束下，使无人机- bs的数量最小化。此外，在没有地面基站的情况下，我们还致力于构建无人机- bs网络的双连通拓扑结构。然而，作为几何磁盘覆盖（GDC）问题的一个特例，公式化问题是np困难的。为了解决这一问题，我们提出了一种启发式算法，即改进QoS-Prior Coverage and bi-Connectivity (IQP2C)，该算法分别解决用户覆盖和双连通拓扑结构子问题。首先，IQP2C以最少的覆盖无人机为用户提供全覆盖。然后，我们提出了一种基于高度簇的方法，从二维Hamilton循环扩展到构建UAV-BS网络的双连通性。仿真结果验证了IQP2C在满足不同QoS需求和构建容错拓扑方面的有效性。此外，IQP2C在用户覆盖的最小化UAV-BSs数量、双向连接的最小化UAV-BSs数量以及运行时间方面优于其他基准。

{"title":"Fault-tolerant 3-D topology construction of UAV-BSs for full coverage of users with different QoS demands","authors":"Xijian Luo , Jun Xie , Liqin Xiong , Yaqun Liu , Yuan He","doi":"10.1016/j.jnca.2025.104332","DOIUrl":"10.1016/j.jnca.2025.104332","url":null,"abstract":"<div><div>Deploying Unmanned aerial vehicle mounted base stations (UAV-BSs) in post-disaster areas or battlefields, where the ground infrastructures are missing or destroyed, can quickly restore communication coverage. Due to the unstable and hostile properties of the environments, the ability to maintain the connectivity of the UAV-BSs network should be considered. In this paper, we study the deployment of UAV-BSs to provide full coverage for users with different quality of service (QoS) demands. The objective is to minimize the number of UAV-BSs under the constraints of user demands and UAV-BS service abilities. Besides, in absence of ground base stations, we also aim to construct a bi-connected topology for the UAV-BS network. However, the formulated problem, as a special instance of the geometric disk cover (GDC) problem, is NP-hard. To tackle this problem, we propose a heuristic algorithm, named Improved QoS-Prior Coverage and bi-Connectivity (IQP2C), by separately solving the user coverage and bi-connected topology construction subproblems. Firstly, IQP2C provides full coverage for users with minimum covering UAVs. Then, we propose an altitude-cluster-based method extending from the 2-D Hamilton cycle to construct bi-connectivity for the UAV-BS network. Simulation results validate the effectiveness of IQP2C in meeting different QoS demands and constructing fault-tolerant topology. Moreover, IQP2C outperforms other baselines in terms of minimized number of UAV-BSs for user coverage, minimized number of UAV-BSs for bi-connectivity as well as running time.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104332"},"PeriodicalIF":8.0,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145094146","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}

引用次数: 0

Artificial intelligence-enhanced zero-knowledge proofs for privacy-preserving digital forensics in cloud environments 用于云环境中保护隐私的数字取证的人工智能增强零知识证明

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-17 DOI: 10.1016/j.jnca.2025.104331

Khizar Hameed , Faiqa Maqsood , Zhenfei Wang

This paper proposed an Artificial Intelligence (AI) enhanced Zero Knowledge Proofs (ZKPs) based comprehensive framework used to improve security, privacy, scalability and efficiency in forensic investigations for the multi-cloud environment, a growing concern for cybersecurity and digital forensics domains. With the growing invulnerability of data storage and inefficient processing in cloud computing landscapes, forensic investigations confront privacy preservation, data integrity, and interoperability issues amongst various cloud providers. Despite existing frameworks, there are few adaptive solutions that holistically solve these challenges. To address such issues and challenges, we propose a suite of frameworks, including an Adaptive Multi-Cloud Forensic Integration Framework (A-MCFIF), Multi-Factor Access Control Framework (MACF), Adaptive ZKP Optimization Framework (AZOF), and Privacy Enhanced Data Security Framework (PDSF) to bridge this gap. Incorporating AI-enhanced ZKP and Multi-Factor Authentication (MFA), these frameworks secure data and improve the efficiency of proof generation and verification while meeting privacy regulations such as the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA). Our extensive evaluation of the proposed framework included computing efficiency, memory consumption, data handling efficiency, scalability, overall performance, and cost-effectiveness. We also analyse verification latency to assess the framework’s real-time processing capabilities, which overcome existing solutions. Furthermore, our research includes cloud-specific threat models such as insider threats and data breaches and shows the benefits of the proposed framework for counteracting these risks by proving mathematical and empirical security against privacy breaches. Finally, we bring new insights and contribute to the development of secure, privacy-compliant, and efficient forensic processes, which are elaborated as a comprehensive solution for more reconstructive forensic initiatives in increasingly sophisticated cloud environments.

本文提出了一种基于人工智能（AI）增强的零知识证明（ZKPs）的综合框架，用于提高多云环境下取证调查的安全性、隐私性、可扩展性和效率，这是网络安全和数字取证领域日益关注的问题。随着云计算环境中数据存储的无懈可击性和低效率处理的增加，取证调查面临着各种云提供商之间的隐私保护、数据完整性和互操作性问题。尽管已有框架，但很少有自适应解决方案能够全面解决这些挑战。为了解决这些问题和挑战，我们提出了一套框架，包括自适应多云取证集成框架（a - mcfif）、多因素访问控制框架（MACF）、自适应ZKP优化框架（AZOF）和隐私增强数据安全框架（PDSF），以弥合这一差距。这些框架结合了人工智能增强的ZKP和多因素身份验证（MFA），可以保护数据，提高证据生成和验证的效率，同时满足隐私法规，如《通用数据保护条例》（GDPR）和《健康保险流通与责任法案》（HIPAA）。我们对提议的框架进行了广泛的评估，包括计算效率、内存消耗、数据处理效率、可伸缩性、整体性能和成本效益。我们还分析了验证延迟，以评估框架的实时处理能力，这克服了现有的解决方案。此外，我们的研究包括特定于云计算的威胁模型，如内部威胁和数据泄露，并通过证明针对隐私泄露的数学和经验安全性，展示了所提议的框架在抵消这些风险方面的好处。最后，我们带来了新的见解，并为开发安全、符合隐私和高效的取证流程做出了贡献，这些流程被阐述为在日益复杂的云环境中更具重建性的取证计划的综合解决方案。

{"title":"Artificial intelligence-enhanced zero-knowledge proofs for privacy-preserving digital forensics in cloud environments","authors":"Khizar Hameed , Faiqa Maqsood , Zhenfei Wang","doi":"10.1016/j.jnca.2025.104331","DOIUrl":"10.1016/j.jnca.2025.104331","url":null,"abstract":"<div><div>This paper proposed an Artificial Intelligence (AI) enhanced Zero Knowledge Proofs (ZKPs) based comprehensive framework used to improve security, privacy, scalability and efficiency in forensic investigations for the multi-cloud environment, a growing concern for cybersecurity and digital forensics domains. With the growing invulnerability of data storage and inefficient processing in cloud computing landscapes, forensic investigations confront privacy preservation, data integrity, and interoperability issues amongst various cloud providers. Despite existing frameworks, there are few adaptive solutions that holistically solve these challenges. To address such issues and challenges, we propose a suite of frameworks, including an Adaptive Multi-Cloud Forensic Integration Framework (A-MCFIF), Multi-Factor Access Control Framework (MACF), Adaptive ZKP Optimization Framework (AZOF), and Privacy Enhanced Data Security Framework (PDSF) to bridge this gap. Incorporating AI-enhanced ZKP and Multi-Factor Authentication (MFA), these frameworks secure data and improve the efficiency of proof generation and verification while meeting privacy regulations such as the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA). Our extensive evaluation of the proposed framework included computing efficiency, memory consumption, data handling efficiency, scalability, overall performance, and cost-effectiveness. We also analyse verification latency to assess the framework’s real-time processing capabilities, which overcome existing solutions. Furthermore, our research includes cloud-specific threat models such as insider threats and data breaches and shows the benefits of the proposed framework for counteracting these risks by proving mathematical and empirical security against privacy breaches. Finally, we bring new insights and contribute to the development of secure, privacy-compliant, and efficient forensic processes, which are elaborated as a comprehensive solution for more reconstructive forensic initiatives in increasingly sophisticated cloud environments.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104331"},"PeriodicalIF":8.0,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145121250","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}

引用次数: 0

FlowTracker: A refined and versatile data plane measurement approach FlowTracker：一种精炼和通用的数据平面测量方法

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-16 DOI: 10.1016/j.jnca.2025.104334

Xinyue Jiang , Chunming Wu , Zhengyan Zhou , Di Wang , Dezhang Kong , Muhammad Khurram Khan , Xuan Liu

To acquire per-hop flow level information, existing works have made significant contributions to offloading network measurement onto data center switches. Despite this, they still pose challenges due to increasingly complex measurement tasks and massive network traffic. In this paper, we introduce FlowTracker, a flow measurement primitive in the data plane. Our key innovation is a hash-based data structure with constant size and collision resolution, which allows fine-grained and real-time monitoring of various flow statistics. We have fully implemented a FlowTracker prototype on a testbed and used real-world packet traces to evaluate its performance. The results demonstrate FlowTracker’s efficiency under different measurement tasks. For example, with

\sim

0.5 MB of memory, FlowTracker can accurately estimate 98% heavy hitter out of 25K flows, with an average relative error of 1.28%. It also achieves 92.27% higher accuracy in packet delay estimation and 121.83% higher flow set coverage compared to competitors with only 64 KB of memory. Furthermore, FlowTracker imposes minimal overhead, requiring just

\sim

0.04% extra bandwidth for large-scale network processing. With these capabilities, FlowTracker can provide network operators with deep insights and efficient flow control of their networks.

为了获取每跳流量级别信息，现有的工作对将网络测量转移到数据中心交换机上做出了重大贡献。尽管如此，由于日益复杂的测量任务和庞大的网络流量，它们仍然带来了挑战。本文介绍了数据平面上的流量测量原语FlowTracker。我们的关键创新是基于哈希的数据结构，具有恒定的大小和冲突分辨率，可以对各种流量统计进行细粒度和实时监控。我们已经在测试平台上完全实现了FlowTracker原型，并使用真实的数据包跟踪来评估其性能。结果证明了FlowTracker在不同测量任务下的效率。例如，使用~ 0.5 MB的内存，FlowTracker可以准确地估计出25K流中98%的重磅攻击，平均相对误差为1.28%。与只有64 KB内存的竞争对手相比，它在数据包延迟估计方面的准确率提高了92.27%，流集覆盖率提高了121.83%。此外，FlowTracker施加最小的开销，只需要~ 0.04%的额外带宽用于大规模网络处理。有了这些功能，FlowTracker可以为网络运营商提供深入的见解和有效的网络流量控制。

{"title":"FlowTracker: A refined and versatile data plane measurement approach","authors":"Xinyue Jiang , Chunming Wu , Zhengyan Zhou , Di Wang , Dezhang Kong , Muhammad Khurram Khan , Xuan Liu","doi":"10.1016/j.jnca.2025.104334","DOIUrl":"10.1016/j.jnca.2025.104334","url":null,"abstract":"<div><div>To acquire per-hop flow level information, existing works have made significant contributions to offloading network measurement onto data center switches. Despite this, they still pose challenges due to increasingly complex measurement tasks and massive network traffic. In this paper, we introduce FlowTracker, a flow measurement primitive in the data plane. Our key innovation is a hash-based data structure with constant size and collision resolution, which allows fine-grained and real-time monitoring of various flow statistics. We have fully implemented a FlowTracker prototype on a testbed and used real-world packet traces to evaluate its performance. The results demonstrate FlowTracker’s efficiency under different measurement tasks. For example, with <span><math><mo>∼</mo></math></span>0.5 MB of memory, FlowTracker can accurately estimate 98% heavy hitter out of 25K flows, with an average relative error of 1.28%. It also achieves 92.27% higher accuracy in packet delay estimation and 121.83% higher flow set coverage compared to competitors with only 64 KB of memory. Furthermore, FlowTracker imposes minimal overhead, requiring just <span><math><mo>∼</mo></math></span>0.04% extra bandwidth for large-scale network processing. With these capabilities, FlowTracker can provide network operators with deep insights and efficient flow control of their networks.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"244 ","pages":"Article 104334"},"PeriodicalIF":8.0,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145134833","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}

引用次数: 0

Secure and efficient data collaboration in cloud computing: Flexible delegation via hierarchical attribute-based signature 云计算中安全高效的数据协作：通过分层属性签名的灵活委托

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-16 DOI: 10.1016/j.jnca.2025.104328

Wenrui Jiang, Yongjian Liao, Qishan Gao, Han Xu, Hongwei Wang

Data collaboration allows multiple parties to jointly share and modify data stored in the cloud server. As unauthorized users may create or modify the shared data as they want by tampering with requests sent by authorized users to replace them with what the unauthorized users want to send, secure data collaboration in cloud computing requires data integrity protection of requests and precise privilege verification of users. However, while maintaining data integrity, it is difficult for current signature schemes to achieve the following demands: fine-grained access control, high scalability, a flexible and controllable hierarchical delegation mechanism, and efficient signing and verification. Therefore, we designed a scalable and flexible hierarchical attribute-based signature (HABS) model and proposed a signing policy HABS construction using the linear secret sharing scheme to construct an access structure. Furthermore, we proved the unforgeability of our HABS scheme in the standard model. We also analyzed and tested the performance of our HABS scheme and related scheme, and we found that our scheme has less signing computation consumption in large-scale systems with complex policies. Finally, we provided a specified application scenario of HABS used in data collaboration based on cloud computing.

数据协作允许多方共同共享和修改存储在云服务器中的数据。由于未经授权的用户可以通过篡改授权用户发送的请求来随意创建或修改共享数据，从而将其替换为未经授权用户想要发送的内容，因此云计算中的安全数据协作需要对请求进行数据完整性保护，并对用户进行精确的权限验证。然而，目前的签名方案在保证数据完整性的同时，难以实现细粒度的访问控制、高可扩展性、灵活可控的分级授权机制、高效的签名和验证等要求。为此，我们设计了一种可扩展、灵活的分层属性签名模型，并提出了一种基于线性秘密共享方案构建访问结构的签名策略HABS构造方法。此外，我们还证明了我们的HABS方案在标准模型中的不可伪造性。对HABS方案和相关方案的性能进行了分析和测试，发现该方案在具有复杂策略的大型系统中签名计算消耗较少。最后，给出了HABS在基于云计算的数据协作中的具体应用场景。

{"title":"Secure and efficient data collaboration in cloud computing: Flexible delegation via hierarchical attribute-based signature","authors":"Wenrui Jiang, Yongjian Liao, Qishan Gao, Han Xu, Hongwei Wang","doi":"10.1016/j.jnca.2025.104328","DOIUrl":"10.1016/j.jnca.2025.104328","url":null,"abstract":"<div><div>Data collaboration allows multiple parties to jointly share and modify data stored in the cloud server. As unauthorized users may create or modify the shared data as they want by tampering with requests sent by authorized users to replace them with what the unauthorized users want to send, secure data collaboration in cloud computing requires data integrity protection of requests and precise privilege verification of users. However, while maintaining data integrity, it is difficult for current signature schemes to achieve the following demands: fine-grained access control, high scalability, a flexible and controllable hierarchical delegation mechanism, and efficient signing and verification. Therefore, we designed a scalable and flexible hierarchical attribute-based signature (HABS) model and proposed a signing policy HABS construction using the linear secret sharing scheme to construct an access structure. Furthermore, we proved the unforgeability of our HABS scheme in the standard model. We also analyzed and tested the performance of our HABS scheme and related scheme, and we found that our scheme has less signing computation consumption in large-scale systems with complex policies. Finally, we provided a specified application scenario of HABS used in data collaboration based on cloud computing.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104328"},"PeriodicalIF":8.0,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145094149","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}

引用次数: 0

C-PFL: A committee-based personalized federated learning framework C-PFL：基于委员会的个性化联邦学习框架

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-16 DOI: 10.1016/j.jnca.2025.104327

Lifan Pan , Hao Guo , Wanxin Li

Federated Learning (FL) is an emerging machine learning paradigm that enables multiple parties to train a shared model while preserving data privacy collaboratively. However, malicious clients pose a significant threat to FL systems. This interference not only deteriorates model performance but also exacerbates the unfairness of the global model caused by data heterogeneity, leading to inconsistent performance across clients. We propose C-PFL, a committee-based personalized FL framework that improves both robustness and personalization. In contrast to prior approaches such as FedProto (which relies on the exchange of class prototypes), Ditto (which employs regularization between global and local models), and FedBABU (which freezes the classifier head during federated training), C-PFL introduces two principal innovations. C-PFL adopts a split-model design, updating only a shared backbone during global training while fine-tuning a personalized head locally. A dynamic committee of high-contribution clients validates submitted updates without public data, filtering low-quality or adversarial contributions before aggregation. Experiments on MNIST, Fashion-MNIST, CIFAR-10, CIFAR-100, and AGNews show that C-PFL outperforms six state-of-the-art personalized FL baselines by up to 2.89% in non-adversarial settings, and by as much as 6.96% under 40% malicious clients. These results demonstrate C-PFL’s ability to sustain high accuracy and stability across diverse non-IID scenarios, even with significant adversarial participation.

联邦学习（FL）是一种新兴的机器学习范式，它使多方能够在协作保护数据隐私的同时训练共享模型。然而，恶意客户端对FL系统构成了重大威胁。这种干扰不仅会降低模型性能，还会加剧由于数据异构而导致的全局模型的不公平性，从而导致客户机之间的性能不一致。我们提出了C-PFL，一个基于委员会的个性化FL框架，提高了鲁棒性和个性化。与之前的方法，如FedProto（依赖于类原型的交换）、Ditto（在全局和局部模型之间使用正则化）和FedBABU（在联邦训练期间冻结分类器头部）相比，C-PFL引入了两个主要的创新。C-PFL采用分体式设计，在全局训练时只更新共享主干，而局部微调个性化头部。一个由高贡献客户端组成的动态委员会在没有公共数据的情况下验证提交的更新，在聚合之前过滤低质量或对抗性的贡献。在MNIST、时尚-MNIST、CIFAR-10、CIFAR-100和AGNews上进行的实验表明，C-PFL在非对抗性环境下比六个最先进的个性化FL基线高出2.89%，在40%恶意客户端下高出6.96%。这些结果证明了C-PFL能够在不同的非iid场景中保持高精度和稳定性，即使有明显的对抗参与。

{"title":"C-PFL: A committee-based personalized federated learning framework","authors":"Lifan Pan , Hao Guo , Wanxin Li","doi":"10.1016/j.jnca.2025.104327","DOIUrl":"10.1016/j.jnca.2025.104327","url":null,"abstract":"<div><div>Federated Learning (FL) is an emerging machine learning paradigm that enables multiple parties to train a shared model while preserving data privacy collaboratively. However, malicious clients pose a significant threat to FL systems. This interference not only deteriorates model performance but also exacerbates the unfairness of the global model caused by data heterogeneity, leading to inconsistent performance across clients. We propose C-PFL, a committee-based personalized FL framework that improves both robustness and personalization. In contrast to prior approaches such as FedProto (which relies on the exchange of class prototypes), Ditto (which employs regularization between global and local models), and FedBABU (which freezes the classifier head during federated training), C-PFL introduces two principal innovations. C-PFL adopts a split-model design, updating only a shared backbone during global training while fine-tuning a personalized head locally. A dynamic committee of high-contribution clients validates submitted updates without public data, filtering low-quality or adversarial contributions before aggregation. Experiments on MNIST, Fashion-MNIST, CIFAR-10, CIFAR-100, and AGNews show that C-PFL outperforms six state-of-the-art personalized FL baselines by up to 2.89% in non-adversarial settings, and by as much as 6.96% under 40% malicious clients. These results demonstrate C-PFL’s ability to sustain high accuracy and stability across diverse non-IID scenarios, even with significant adversarial participation.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104327"},"PeriodicalIF":8.0,"publicationDate":"2025-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145094150","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}

引用次数: 0

Hybrid connectivity-oriented efficient shielding for robustness enhancement in large-scale networks 面向混合连接的大规模网络鲁棒性增强有效屏蔽

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-14 DOI: 10.1016/j.jnca.2025.104325

Wei Wei , Jie Huang , Qinghui Zhang , Tao Ma , Peng Li

Network infrastructure protection is critical for ensuring robustness against attacks and failures, yet existing approaches fundamentally limit their scope by addressing either node or edge vulnerabilities in isolation — an unrealistic assumption given real-world scenarios where both element types may fail simultaneously. Our work makes three key advances beyond current state-of-the-art: First, we introduce the novel concept of hybrid connectivity as a unified robustness metric that properly accounts for concurrent node-edge failures, demonstrating through theoretical analysis that traditional single-element metrics require prohibitively high connectivity thresholds. Second, we develop the first practical solution for large-scale networks via our hybrid cut-tree mapping algorithm, which employs an extended node cut formulation with dynamic programming to identify all vulnerable node-edge combinations in linear time — a dramatic complexity reduction from the exponential scaling of existing linear programming methods. Third, we prove and exploit a fundamental structural property that shielding any edge spanning tree plus leaf edges guarantees target hybrid connectivity, enabling our edge spanning tree algorithm to deliver near-optimal solutions at unprecedented scale. Experimental validation confirms our approach maintains 100% protection effectiveness (with no more than 6% cost overhead versus optimal) in small graphs while achieving 99.9% protection coverage in large-scale networks — outperforming all existing heuristics in protection cost while providing a

1 0^{5}

times speedup over traditional methods.

网络基础设施保护对于确保对攻击和故障的稳健性至关重要，但现有的方法从根本上限制了它们的范围，只能孤立地解决节点或边缘漏洞——考虑到两种元素类型可能同时失效的现实场景，这是一个不切实际的假设。我们的工作取得了超越当前技术水平的三个关键进展：首先，我们引入了混合连接的新概念，将其作为统一的鲁棒性指标，适当地解释并发节点边缘故障，通过理论分析证明传统的单元素指标需要过高的连接阈值。其次，我们通过混合切割树映射算法开发了大规模网络的第一个实用解决方案，该算法采用扩展的节点切割公式和动态规划来识别线性时间内所有脆弱的节点-边缘组合-从现有线性规划方法的指数缩放中显着降低了复杂性。第三，我们证明并利用了一种基本的结构特性，即屏蔽任何边生成树和叶边，保证目标混合连通性，使我们的边生成树算法能够以前所未有的规模提供接近最优的解决方案。实验验证证实，我们的方法在小图形中保持100%的保护有效性（与最优相比，成本开销不超过6%），同时在大规模网络中实现99.9%的保护覆盖率——在保护成本方面优于所有现有的启发式方法，同时提供比传统方法105倍的加速。

{"title":"Hybrid connectivity-oriented efficient shielding for robustness enhancement in large-scale networks","authors":"Wei Wei , Jie Huang , Qinghui Zhang , Tao Ma , Peng Li","doi":"10.1016/j.jnca.2025.104325","DOIUrl":"10.1016/j.jnca.2025.104325","url":null,"abstract":"<div><div>Network infrastructure protection is critical for ensuring robustness against attacks and failures, yet existing approaches fundamentally limit their scope by addressing either node or edge vulnerabilities in isolation — an unrealistic assumption given real-world scenarios where both element types may fail simultaneously. Our work makes three key advances beyond current state-of-the-art: First, we introduce the novel concept of hybrid connectivity as a unified robustness metric that properly accounts for concurrent node-edge failures, demonstrating through theoretical analysis that traditional single-element metrics require prohibitively high connectivity thresholds. Second, we develop the first practical solution for large-scale networks via our hybrid cut-tree mapping algorithm, which employs an extended node cut formulation with dynamic programming to identify all vulnerable node-edge combinations in linear time — a dramatic complexity reduction from the exponential scaling of existing linear programming methods. Third, we prove and exploit a fundamental structural property that shielding any edge spanning tree plus leaf edges guarantees target hybrid connectivity, enabling our edge spanning tree algorithm to deliver near-optimal solutions at unprecedented scale. Experimental validation confirms our approach maintains 100% protection effectiveness (with no more than 6% cost overhead versus optimal) in small graphs while achieving 99.9% protection coverage in large-scale networks — outperforming all existing heuristics in protection cost while providing a <span><math><mrow><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>5</mn></mrow></msup></mrow></math></span> times speedup over traditional methods.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104325"},"PeriodicalIF":8.0,"publicationDate":"2025-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145093955","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}

引用次数: 0

5G slicing under the hood: An in-depth analysis of 5G RAN features and configurations 5G切片：5G RAN特性和配置的深入分析

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-13 DOI: 10.1016/j.jnca.2025.104298

André Perdigão, José Quevedo, Rui L. Aguiar

There has been extensive discussion on the benefits and improvements that 5G networks can bring to industry operations, particularly with network slicing. However, to fully realize network slices, it is essential to thoroughly understand the mechanisms available within a 5G network that can be used to adapt network performance. This paper surveys and describes existing 5G network configurations and assesses the performance impact of several configurations using a real-world commercial standalone (SA) 5G network, bringing the challenges between purely theoretical mathematical models into realizations with existing equipment. The paper discusses how these features impact communication performance according to industrial requirements.

The survey describes and demonstrates the performance impact of various 5G configurations, enabling readers to understand the capabilities of current 5G networks and learn how to leverage 5G technology to enhance industrial operations. This knowledge is also crucial to fully realize network slices tailored to industrial requirements.

关于5G网络可以给行业运营带来的好处和改进，特别是网络切片，人们进行了广泛的讨论。然而，要完全实现网络切片，必须彻底了解5G网络中可用的机制，这些机制可用于调整网络性能。本文调查和描述了现有的5G网络配置，并使用实际商用独立（SA） 5G网络评估了几种配置对性能的影响，将纯理论数学模型之间的挑战转化为现有设备的实现。本文根据工业需求，讨论了这些特性对通信性能的影响。

{"title":"5G slicing under the hood: An in-depth analysis of 5G RAN features and configurations","authors":"André Perdigão, José Quevedo, Rui L. Aguiar","doi":"10.1016/j.jnca.2025.104298","DOIUrl":"10.1016/j.jnca.2025.104298","url":null,"abstract":"<div><div>There has been extensive discussion on the benefits and improvements that 5G networks can bring to industry operations, particularly with network slicing. However, to fully realize network slices, it is essential to thoroughly understand the mechanisms available within a 5G network that can be used to adapt network performance. This paper surveys and describes existing 5G network configurations and assesses the performance impact of several configurations using a real-world commercial standalone (SA) 5G network, bringing the challenges between purely theoretical mathematical models into realizations with existing equipment. The paper discusses how these features impact communication performance according to industrial requirements.</div><div>The survey describes and demonstrates the performance impact of various 5G configurations, enabling readers to understand the capabilities of current 5G networks and learn how to leverage 5G technology to enhance industrial operations. This knowledge is also crucial to fully realize network slices tailored to industrial requirements.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104298"},"PeriodicalIF":8.0,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145059766","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}

引用次数: 0

Reinforcement learning based mobile charging sequence scheduling algorithm for optimal stochastic event detection in wireless rechargeable sensor networks 基于强化学习的无线可充电传感器网络随机事件最优检测移动充电序列调度算法

IF 8 2区计算机科学 Q1 COMPUTER SCIENCE, HARDWARE & ARCHITECTURE

Journal of Network and Computer Applications

Pub Date : 2025-09-13 DOI: 10.1016/j.jnca.2025.104301

Jinglin Li , Haoran Wang , Sen Zhang , Peng-Yong Kong , Wendong Xiao

Mobile charging provides a new way for energy replenishment in Wireless Rechargeable Sensor Network (WRSN), where the Mobile Charger (MC) is employed for charging sensor nodes sequentially according to the mobile charging sequence scheduling result. Event detection is an essential application of WRSN, but when the events occur stochastically, Mobile Charging Sequence Scheduling for Optimal Stochastic Event Detection (MCSS-OSED) is difficult and challenging, and the non-deterministic detection property of the sensor makes MCSS-OSED complicated further. This paper proposes a novel Multistage Exploration Q-learning Algorithm (MEQA) for MCSS-OSED based on reinforcement learning. In MEQA, MC is taken as the agent to explore the space of the mobile charging sequences via the interactions with the environment for the optimal Quality of Event Detection (QED) evaluated by both considering the sensing probability of the sensor and the probability that events may occur in the monitoring region. Particularly, a new multistage exploration policy is designed for MC to improve the exploration efficiency by selecting the current suboptimal actions with a certain probability, and a novel reward function is presented to evaluate the MC charging action according to the real-time detection contribution of the sensor. Simulation results show that MEQA is efficient for MCSS-OSED and superior to the existing classical algorithms.

移动充电为无线可充电传感器网络（WRSN）提供了一种新的能量补充方式，移动充电器（MC）根据移动充电顺序调度结果对传感器节点进行顺序充电。事件检测是WRSN的重要应用，但当事件随机发生时，针对最优随机事件检测的移动充电序列调度（MCSS-OSED）是一个难点和挑战，而传感器的不确定性检测特性使MCSS-OSED更加复杂。提出了一种基于强化学习的MCSS-OSED多阶段探索q学习算法。在MEQA中，MC作为智能体，通过与环境的相互作用来探索移动充电序列的空间，通过考虑传感器的感知概率和监测区域可能发生事件的概率来评估最优事件检测质量（QED）。设计了一种新的多阶段探索策略，通过选择具有一定概率的当前次优行为来提高MC的探索效率，并提出了一种新的奖励函数，根据传感器的实时检测贡献来评估MC的收费行为。仿真结果表明，MEQA算法在MCSS-OSED中是有效的，优于现有的经典算法。

{"title":"Reinforcement learning based mobile charging sequence scheduling algorithm for optimal stochastic event detection in wireless rechargeable sensor networks","authors":"Jinglin Li , Haoran Wang , Sen Zhang , Peng-Yong Kong , Wendong Xiao","doi":"10.1016/j.jnca.2025.104301","DOIUrl":"10.1016/j.jnca.2025.104301","url":null,"abstract":"<div><div>Mobile charging provides a new way for energy replenishment in Wireless Rechargeable Sensor Network (WRSN), where the Mobile Charger (MC) is employed for charging sensor nodes sequentially according to the mobile charging sequence scheduling result. Event detection is an essential application of WRSN, but when the events occur stochastically, Mobile Charging Sequence Scheduling for Optimal Stochastic Event Detection (MCSS-OSED) is difficult and challenging, and the non-deterministic detection property of the sensor makes MCSS-OSED complicated further. This paper proposes a novel Multistage Exploration Q-learning Algorithm (MEQA) for MCSS-OSED based on reinforcement learning. In MEQA, MC is taken as the agent to explore the space of the mobile charging sequences via the interactions with the environment for the optimal Quality of Event Detection (QED) evaluated by both considering the sensing probability of the sensor and the probability that events may occur in the monitoring region. Particularly, a new multistage exploration policy is designed for MC to improve the exploration efficiency by selecting the current suboptimal actions with a certain probability, and a novel reward function is presented to evaluate the MC charging action according to the real-time detection contribution of the sensor. Simulation results show that MEQA is efficient for MCSS-OSED and superior to the existing classical algorithms.</div></div>","PeriodicalId":54784,"journal":{"name":"Journal of Network and Computer Applications","volume":"243 ","pages":"Article 104301"},"PeriodicalIF":8.0,"publicationDate":"2025-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145120317","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}

引用次数: 0