检索结果-内蒙古大学图书馆

Evolutionary Adaptation Mechanism for Edge Caching Under Propagation Dynamics

IEEE INTERNET OF THINGS JOURNAL 2024年第9期11卷 15179-15192页

作者： Fan, Shaoshuai Gao, Hanlin Tian, Hui Yang, Shiyu Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China China Telecom Corp Ltd Dalian 110015 Liaoning Peoples R China China Construct Bank Beijing 100053 Peoples R China

In mobile-edge network, adaptive cache placement and resource allocation are vital for meeting the dynamic requirements of users. The current network cache performance has not been fully implemented, due to the lack of comprehensive knowledge of content popularity dynamics and the impact of network resource deployment on the dynamics. In this article, an epidemic model is applied to characterize the content propagation among users. Based on this model, a joint optimization problem of cache placement and base station resource allocation is formulated, maximizing the satisfaction rate of users. The joint optimization problem is decomposed into two subproblems: 1) caching placement and 2) resource allocation. First, the cache placement strategy is determined based on the dynamic propagation model. Second, considering the impact of network decisions on content propagation, a belief propagation-based iterative algorithm is proposed for dynamic resource allocation. Finally, simulation results reveal that the proposed joint optimization algorithm outperforms the benchmark algorithms.

关键词： Resource management Heuristic algorithms Dynamic scheduling Epidemics Adaptation models User experience Optimization Belief propagation edge caching epidemic model resource allocation

来源：评论

学校读者我要写书评

暂无评论

Flexible and efficient network slicing scheme based on resource awareness and traffic prediction for Fi-Wi access network

引用

OPTICAL FIBER technology 2025年 92卷

作者： Li, Kun Tian, Qinghua Xin, Xiangjun Zhang, Xiao Li, Zuxian Wang, Fu Yu, Wensheng Tian, Feng Yao, Haipeng Zhang, Qi Beijing Univ Posts & Telecommun BUPT State Key Lab Informat Photon & Opt Commun Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun Beijing Key Lab Space Ground Interconnect & Conver Beijing 100876 Peoples R China Beijing Inst Technol Sch Informat & Elect Beijing 100081 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

Optical access networks are widely recognized as a pivotal technology for addressing the last-mile connectivity challenges. However, the continuous emergence of novel applications within the industrial Internet introduces unprecedented challenges to these networks. This paper proposes a resource-aware dynamic slicing scheme for Fi-Wi access networks. This proposed scheme adheres to the ISA-95 standard and seamlessly integrates network slicing technology into optical access networks. The scheme uses a resource-aware model to select the optimal path and reduce processing latency. LSTM-Attention is introduced for traffic prediction, reducing the delay in slice adjustments caused by large-scale traffic. Simulation results show that this scheme can efficiently select the optimal path, reserve more resources for services, and has a lower bandwidth blocking rate compared to traditional DES and LSTM methods.

关键词： Fi-Wi access network Industrial internet Software defined network Resource management LSTM-Attention

来源：评论

学校读者我要写书评

暂无评论

An Experimental Analysis of Channel Similarity in the UMa Scenario for 6G FR3 Communications 7

An Experimental Analysis of Channel Similarity in the UMa Sc...

引用

7th IEEE International Conference on Electronic Information and Communication technology, ICEICT 2024

作者： Chen, Ke Tang, Pan Jiang, Tao Xu, Huixin Nie, Gaofeng Tian, Lei Zhang, Jianhua Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology Beijing100876 China China Mobile Research Institution Beijing100053 China

ISBN: (纸本)9798350384437

In the development of the sixth generation (6G) wireless communication, FR3 (7-24 GHz) has gradually become the focus band due to its advantages of wide coverage and high capacity. However, the span of the FR3 is large, and the channel characteristics within the band have not been fully studied. In this paper, we conduct channel measurements campaigns in Urban Macro (UMa) scenario at FR3 using the channel sounder, and analyse the channel similarity. Firstly, channel measurements are conducted in UMa scenario at 11-15 GHz using a time domain channel measurement platform, with three receiver antennas of different beamwidths. Subsequently, the channel impulse response (CIR) similarity in frequency domain and beam domain is analyzed using pearson correlation coefficient (PCC) and temporal power delay distribution correlation coefficient (TPCC) as metrics. Finally, it is found that as the channel frequency interval increases, there is a trend of decreasing channel similarity. Additionally, great differences in antenna beamwidths also lead to lower channel similarity. © 2024 IEEE.

关键词： Impulse response

来源：评论

学校读者我要写书评

暂无评论

Modeling and analysis of computation offloading in NOMA-based fog radio access networks

引用

WIRELESS NETWORKS 2024年第3期30卷 1305-1319页

作者： Lin, Lixia Yang, Zhicheng Dong, Zhihong Beijing Inst Graphic Commun Sch Informat Engn 1 Xinghua St Beijing 102600 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol 10 Xitucheng Rd Beijing 100876 Peoples R China

The hierarchical local-fog-cloud computing-enabled paradigm applied in fog radio access networks (F-RANs) has been considered as a promising architecture to cope with the surge of mobile intelligent applications. Meanwhile, non-orthogonal multiple access (NOMA) technology with superior spectrum efficiency has been widely utilized to provide transmission services for offloading tasks in F-RANs. A comprehensive understanding of the computation offloading performance of NOMA-based F-RANs is thus essential to provide guidelines for F-RANs planning and resources scheduling. In this paper, we first develop a mathematical framework to analyze the ability of NOMA to support computation offloading in F-RANs. Then, the expression for a lower bound of the average computation offloading probability is provided based on stochastic geometry and order statistics, and the relation between the computation offloading probability and the average task execution latency as well as energy consumption is further derived by queuing theory. Finally, extensive numerical simulations are executed to verify the accuracy of our theoretical results. The numerical results show that the average computation offloading probability is nearly linear with most of key network parameters, such as the coverage radius of cellular networks, the density of fog radio access points (F-APs), the number of users sharing the same spectrum resources, and the task request rate, respectively.

关键词： Fog radio access networks Non-orthogonal multiple access Stochastic geometry Computation offloading probability

来源：评论

学校读者我要写书评

暂无评论

Spatio-Temporal Aware Personalized Federated Learning for Load Forecasting in Power Systems

引用

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS 2025年

作者： Xu, Siya Zou, Jie Zhou, Cheng Wu, Hai Zeng, Zeng Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China NARL Technol Co Ltd Nanjing 211006 Peoples R China State Grid Jiangsu Elect Power Co Ltd Informat Commun Branch Nanjing 210003 Peoples R China

With the development of smart grids and the increasing demand for electric energy consumption, electricity load forecasting has become more and more important in electric energy management. However, the differences in electricity load patterns between different regions lead to data heterogeneity, which may seriously affect the model performance of traditional federated learning methods in electricity load forecasting. Meanwhile, the resource heterogeneity between clients will further decrease the efficiency and the accuracy of forecasting. To this end, we propose a spatio-temporal aware personalized federated learning (PFL) framework for electricity load forecasting to improve the forecasting accuracy and training speed so as to enhance the real-time responsiveness and system stability of the power grid. First, to solve the data heterogeneity, we design a collaborative training domain (CTD) construction method based on spatio-temporal features. Then, on the basis of constructed CTDs, we propose a spatio-temporal convolutional network (STCN)-based layered PFL method to address the resource heterogeneity, which can be divided into personalized layers and generalized layers according to temporal and spatial static features separately. In addition, we design a hierarchical aggregation mechanism and an adaptive edge model aggregation adjustment mechanism to optimize the training process. Experimental results show that the method proposed in this article outperforms other methods in terms of model accuracy and convergence speed.

关键词： \Data heterogeneity electricity load forecasting personalized federated learning (PFL) resource heterogeneity spatio-temporal convolutional network (STCN)

来源：评论

学校读者我要写书评

暂无评论

Fast Reroute Mechanism for Satellite Networks Based on Segment Routing and Dual Timers switching

引用

AEROSPACE 2025年第3期12卷 233-233页

作者： Du, Jinyan Zhang, Ran Hu, Jiangbo Xia, Tian Liu, Jiang China United Network Commun Grp Co Ltd Beijing 100033 Peoples R China Beijing Univ Posts & Telecommun State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Purple Mt Labs Nanjing 211111 Peoples R China

Low-Earth-Orbit (LEO) satellite networks have the advantage of global internet coverage and low latency, and they have enjoyed great success in the past few years. In LEO satellite networks, laser-based inter-satellite links (ISLs) are widely employed to achieve on-board data relay, and further to provide high-capacity backhaul worldwide. However, ISLs are prone to break due to the outage of the ISL capturing, tracking, and aiming systems. Meanwhile, breaks caused by different reasons can last from milliseconds to hours. The hybrid ISL fault leads to the on-board routing protocol to flap frequently, thus causing high routing overheads, low convergence speed, and degraded service consistency. In this work, we propose a hybrid fault detection mechanism to identify transient and long-term ISL outage. Further, for transient link outage, the segment routing-based loop-free backup path is adopted to provide real-time transmission recovery, and precise global route convergence is adopted to restore the long-term routing failure. For the inconsistent routing table switch between the phase from transient to long-term fault, we propose a dual timer mechanism to make sure the path can be smoothly switched without micro-loops. Simulation results validate the feasibility and efficiency of the proposed scheme.

关键词： segment routing fast rerouting dual-timer mechanism backup path design LEO satellite networks

来源：评论

学校读者我要写书评

暂无评论

SOHO-FL: A Fast Reconvergent Intra-Domain Routing Scheme Using Federated Learning

引用

IEEE NETWORK 2024年第1期38卷 272-279页

作者： Cong, Peizhuang Zhang, Yuchao Wang, Wendong Xu, Ke Beijing Univ Posts & Telecommun BUPT State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China Beijing Univ Posts & Telecommun BUPT Beijing 100876 Peoples R China Tsinghua Univ Beijing 100084 Peoples R China

The development of machine learning provides a new paradigm for network optimizations, e.g., reinforcement learning (RL) has brought great improvements in many fields, such as adaptive video streaming, congestion control of TCP. The fundamental mechanism of such RL-based architectures is that the neural network decision model converges to a stable state by continuously interacting with network environment. However, for network routing problem, such RL-based strategies do not work well due to topology change. This is because topological changes would require the existing RL models to be retrained, while these models may stop making routing decisions or provide non-optimal decisions during the slow reconverging process of retraining, seriously affected transmission performance. To solve this problem, we proposed a fast convergent RL-model (SOHO-FL), which can alleviate the performance degradation caused by the slow retraining process by federated learning. The experimental results based on real-world network topologies demonstrate that SOHO-FL outperforms the state-of-the-art algorithms in reconvergence time by 22.3% on average.

关键词： Routing Hidden Markov models Data models Computational modeling Network topology Optimization Federated learning

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection Method Based on Complementary Adversarial Generation Network 14th

Intrusion Detection Method Based on Complementary Adversaria...

引用

14th International Conference on Swarm Intelligence (ICSI)

作者： Li, Lixiang Liu, Yuxuan Peng, Haipeng Beijing Univ Posts & Telecommun Informat Secur Ctr State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

ISBN: (纸本)9783031366246;9783031366253

How to effectively identify the attack behavior in the network environment is an urgent problem in the field of network security. In this paper, an intrusion detection method based on complementary generative adversarial network is proposed. The algorithm is One Class Classification algorithm based on generative adversarial network architecture. The algorithm uses normal data in training, normal and abnormal data in testing phase, and calculates the accuracy rate. In this algorithm, complementary samples are generated by complementary generator and input to discriminator as exception samples for training. The discriminator in this paper uses stacked Non-Symmetric autoencoders to recognize normal samples from training data and abnormal samples generated by generators through training. Experimental results show that the algorithm proposed in this paper is superior to other algorithms of the same kind in performance, with higher accuracy and excellent time performance.

关键词： Intrusion Detection Deep Learning Complementary GAN Auto-Encoder

来源：评论

学校读者我要写书评

暂无评论

Coverage and reliability analysis of multi-IRS-aided distributed space-time block coding in MISO wireless communication systems

引用

PHYSICAL COMMUNICATION 2024年 63卷

作者： Wang, Xudong Feng, Lei Zhou, Fanqin Li, Wenjing BUPT State Key Lab Networking & Switching Technol Beijing 100876 Peoples R China

Intelligent reflecting surface (IRS) can realize information manipulation and signal modulations to perform space-time coding wireless communication. This article introduces a novel model to perform space-time block coding (STBC), where the base station transmits multiple symbols to a destination through multiple IRS that are utilized to adjust the phase shift and establish strong links. To maximize the end-to-end SNR, we propose a graph theory based method to obtain the optimal transmit antenna selection (TAS) scheme. Based the optimal TAS scheme, we derive the statistical characterizations of the sum of independent generalized gamma random variables to obtain end-to-end signal-to-noise ratio expressions. Then, the closed-form expressions of outage probability (OP) and symbol error rate (SER) are formulated respectively to evaluate the coverage and reliability performance in wireless communication systems. Furthermore, the asymptotic expressions for OP and SER at a high transmission power are given, then the diversity order is discussed to observe the performance limit. To verify this theoretical performance analysis, extensive computer simulations are conducted to prove its effectiveness. The simulation results show that the multi-IRS-aided STBC system can attain the same or even better performance as the STBC system based on multiple amplify-and-forward relays using a little number of reflecting elements. More interestingly, the OP and SER of this system decrease more than that of the relays-aided system as transmission power increases.

关键词： Intelligent reflecting surface Space-time block coding Outage probability Symbol error rate Phase shift Relay

来源：评论

学校读者我要写书评

暂无评论

A Transformer-Block-Wise Collaborative Training Mechanism with Hybrid Parallelism Over Heterogeneous Networks

A Transformer-Block-Wise Collaborative Training Mechanism wi...

引用

2025 IEEE Wireless Communications and networking Conference, WCNC 2025

作者： Chen, Jiewei Wang, Jingrong Guo, Shaoyong Hao, Jiakai Qiu, Xuesong Xiong, Zehui Beijing University of Posts and Telecommunications State Key Laboratory of Networking and Switching Technology China State Grid Beijing Information & Communication Company China Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore

ISBN: (纸本)9798350368369

With the rise of AI-Generated Content (AIGC) services in wireless networks, efficient and high-quality distributed training of Large Language Models (LLMs) has become essential for enabling the large-scale application of next generation AI technologies. However, the extensive parameters of LLMs impose significant demands on memory, computing power and communication resources in heterogeneous networks. To efficiently utilize the dispersed network resources, this paper presents a First-Pipeline- Then-Federated Learning (FPTFL) approach with a hybrid parallel scheduling strategy to facilitate the training of Transformer-based LLMs. We propose a block-wise splitting mechanism to partition the Transformer's encoder into distinct segments, which are deployed cross individual devices. The encoder parameters and intermediate smashed data are uploaded to the edge server, where the whole model is updated through federated aggregation. Particularly, we develop a fine-grained computation-efficient method based on pipeline parallelism, enabling the segments to cooperatively train the entire encoder. An optimization problem is formulated to determine the LLM segments and the number of micro-batches under network resource constraints, with the goal of minimizing the total latency of LLM training services. Simulation results demonstrate that our approach enables Transformer-based model training on resource-constrained devices, preserves model performance, and reduces waiting time. © 2025 IEEE.

关键词： Federated learning large language model (LLM) pipeline parallelism training services

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：