Internet of Things Devices (IoTDs) in remote areas connect to the Internet through Low Earth Orbit (LEO) satellites. The transmission of massive IoTDs data volumes leveraging satellites faces a critical challenge, as constrained bandwidth onboard significantly impacts the performance of time-sensitive applications. Moreover, due to the limited number of satellite-ground links and the capacity of satellite access devices, it is challenging to serve all IoTDs within the satellite network coverage. To tackle these challenges, we propose a dual-layer network architecture that integrates Satellite Internet of Things (SIoT) and Unmanned Aerial Vehicles (UAVs), where UAVs serve as transmission units and satellites serve as computing units to overcome the limitations of conventional SIoT. In the dual-layer network architecture, a dual-layer routing problem is formulated to ensure the rapid transmission of tasks, and the problem is further decomposed into UAV layer routing and LEO satellite layer routing. At the UAV layer, we propose a multitask concurrent routing strategy based on task segmentation according to path bandwidth to reduce link load and improve transmission efficiency. At the LEO satellite layer, an integrated computation-transmission routing strategy is introduced to mitigate onboard bandwidth constraints through computational capabilities, thereby reducing transmitted data volume and significantly enhancing transmission efficiency. We transform the problem into a Markov Decision Process (MDP) for each layer and solve the problem using Deep Reinforcement Learning (DRL). To enhance the performance, we introduce improvements to the route algorithm. In the UAV layer, we introduce channel state averaging to reduce algorithmic complexity. In the LEO layer, we employ Prioritized Trajectory Replay (PTR) to improve learning efficiency, while a loss constraint is introduced to enhance training stability. Simulation results demonstrate that the proposed algorithm outperforms other algorithms in terms of convergence performance and overall delay.

At present, most edge servers adopt popularity-based caching strategies, prioritizing the caching of content with the highest overall popularity on user-side edge servers. However, in social network scenarios, user interests and preferences are highly personalized and dynamically changing. This results in existing caching strategies often failing to adjust the cache placement of content in real time according to individual user preferences, leading to suboptimal edge cache hit rates, increased user request response latency, and a decline in quality of service (QoS) for the user experience. To address this issue, we propose a new caching strategy tailored for large-scale social content based on knowledge graph neural network (KGNNC). First, an entity-relation KG is constructed from users' triple data$(\boldsymbol{h}, \boldsymbol{r}, \boldsymbol{t})$on social platforms. Next, a graph convolutional neural network is employed to iteratively aggregate feature information from neighboring nodes and learn vector representations of the nodes. Finally, a reinforcement learning-based algorithm is utilized to determine the optimal caching location for content. Experimental results on multiple public datasets demonstrate, compared with several existing baseline algorithms (least recently used, least frequently used, neural network-based collaborative filtering, KG-DQN, and CAFR), the algorithm proposed in this article achieves a reduction in the average response latency of requests by 38.26%, 34.46%, 13.56%, 6.49%, 4.19% on MovieLens 1M dataset and 30.31%, 27.59%, 14.11%, 5.83%, 4.78% on last FM dataset, respectively. Meanwhile, experiment results demonstrate that the caching hit rate is increased by 24.2%, 25.1%, 14.4%, 6.5%, 3.53% on MovieLens 1M and 39.51%, 39.05%, 23.40%, 10.66%, 8.67% on last FM compared with the four existing baseline algorithms, respectively. These results verify the effectiveness of our algorithm in reducing response latency of user requests and improving caching hit rate of edge servers in social networks.

With the rapid development of 6G and Internet of Vehicles (IoV) technologies, the volume of computation-intensive tasks generated by intelligent vehicles is growing exponentially. Given limited onboard processing capabilities, vehicles increasingly rely on edge servers deployed by service providers (SPs) at roadside units to offload tasks. Vehicle clients can offload the tasks to SPs to mitigate their onboard computation load, while SPs derive economic benefits through the provision of computation resources. However, this interaction introduces a conflict of interest, as vehicles aim to minimize their offloading costs, while SPs seek to maximize revenue. To address this problem, we propose SPOR, a Stackelberg game-based service priority-aware computation offloading and resource pricing scheme in IoV. SPOR is a hierarchical game-theoretic framework in which SPs act as leaders setting prices, while vehicles act as followers determining their offloading strategies. A novel service prioritization function is introduced, incorporating booking price, system load, and reputation to ensure fair and balanced resource allocation. We provide a theoretical proof of the existence and uniqueness of a Nash equilibrium. Extensive experiments on a real-world vehicle edge computing dataset show that SPOR outperforms baseline methods in delay, energy consumption, average load, and task completion rate. Notably, SPOR maintains task completion rates above 97% even under heavy workloads, demonstrating its effectiveness in enhancing system reliability and overall performance.

As the in-network computing (INC) paradigm evolves, efficient scheduling of dependent tasks within complex network systems becomes increasingly crucial. The network needs to handle high-level resource demands while adhering to strict latency requirements. Deterministic delay constraints are particularly critical in applications that rely on directed acyclic graphs (DAGs). To address this challenge, we first propose a deterministic delay-aware task scheduling optimization problem over INC to maximize resource utilization and ensure task acceptance. We accurately establish the complex deterministic delay constraint through traffic arrival and service curves and utilize network calculus for conversion to facilitate solving. Then, we further transform the task optimization problem into MDP and develop a deep reinforcement learning (DRL) algorithm that combines graph neural network (GNN) and delay-aware proximal policy optimization (DPPO) to solve it, called the Deterministic Delay-aware Task Scheduling (DDTS) scheme. It utilizes multilayer GNN to handle task dependencies and applies the DPPO algorithm to introduce deterministic delay penalty factors to evaluate policy operations, achieving optimal task scheduling. The simulation results demonstrate the significant advantages of the DDTS scheme over existing algorithms and task scheduling schemes in terms of task acceptance rate and resource utilization.

The evolution of 6G networks introduces heterogeneous services with stringent computing and latency demands. However, constrained edge resources, intricate task dependencies, and dynamic network fluctuations intensify resource contention, increasing the risk of node faults and service interruption. Current fault-tolerant methodologies lack the necessary adaptability to handle the coupled complexity of task interdependencies and volatile resource states, leading to sub-optimal decisions or excessive system overhead. To address these challenges, this paper innovatively proposes TransDiffuse—an intelligent preemptive service migration framework for 6G edge networks. First, the framework employs a Transformer-GAT hybrid model to capture long-range temporal load dynamics and spatial topological constraints, enabling accurate failure prediction. Second, to navigate the trade-off between migration overhead and service robustness, we devise a diffusion-based decision module. This module efficiently explores the discrete combinatorial solution space to synthesize near-optimal service orchestration. Furthermore, a comprehensive evaluation system is constructed to validate the effectiveness of TransDiffuse. Experiments demonstrate that TransDiffuse reduces energy consumption by 32.4%, decreases task completion time by 25.6%, and improves resource balance by 18.7%, while keeping service violations below 5%. This work achieves joint optimization of energy, delay, and resource efficiency, offering a robust solution for resilient service orchestration in 6G edge networks.

Task scheduling decisions for multimodal large language model (MLLM) applications in multilayer computing power networks present a significant challenge, as they simultaneously balance system delay, carbon emissions, and model accuracy requirements while adapting to network conditions and varying energy availability. Thus, in this paper, we formulate the joint optimization problem of MLLM task scheduling, resource allocation, and green energy utilization to minimize system delay and carbon emissions while meeting accuracy requirements. We propose Green-Aware MAPPO, a novel approach that integrates graph attention networks (GAT) with multi-agent proximal policy optimization (MAPPO) for distributed decision-making in multilayer computing power networks. By modeling the problem as a partially observable Markov decision process (POMDP), our algorithm enables agents to capture complex resource dependencies through relation-specific attention mechanisms while maintaining high performance with limited local observations. Experiments in various network configurations demonstrate that Green-Aware MAPPO significantly outperforms baseline algorithms.

Digital twin (DT), as a key enabling technology for 6G edge intelligence, can establish real-time connections between digital twin objects and physical devices, ensuring real-time synchronization and thereby enhancing the service performance and stability of edge networks. This paper considers the high dynamics of edge networks and combines digital twins with edge networks to construct a three-layer network architecture. Based on the demands for low-latency services and system energy efficiency, we design a network metric: system overhead, to minimize service latency and system energy consumption. To achieve these system objectives, we integrate digital twin technology with multiagent deep reinforcement learning (MADRL), proposing a digital twin-driven multi-agent scheme for green and reliable resource allocation. This approach effectively minimizes system overhead and can adapt well to dynamic changes in terminal devices. Compared with baseline algorithms, it reduces system overhead by at least 7 % while maintaining significant reliable.

Multiview photometric stereo (MVPS) seeks to recover high-fidelity surface shapes and reflectances from images captured under varying views and illuminations. However, existing MVPS methods often require controlled darkroom settings for varying illuminations or overlook the recovery of reflectances and illuminations properties, limiting their applicability in natural illumination scenarios and downstream inverse rendering tasks. In this paper, we propose RotatedMVPS to solve shape and reflectance recovery under rotated natural light, achievable with a practical rotation stage. By ensuring light consistency across different camera and object poses, our method reduces the unknowns associated with complex environment light. Furthermore, we integrate data priors from off-the-shelf learning-based single-view photometric stereo methods into our MVPS framework, significantly enhancing the accuracy of shape and reflectance recovery. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of our approach.

With the development of new power systems, new services in the power data network are constantly emerging and the demand for deterministic transmission is on the rise. The transformation from "best - effort" to "punctual and accurate" has become an urgent issue to be addressed in precision load control services. This paper presents a deterministic routing approach which decouples paths and cycles for wide - area precision load control services within new power systems. The aim is to enhance the scheduling success rate and resource utilization rate of wide area deterministic load control services in new power systems. The method formulates the scheduling success rate and resource utilization rate as optimization objectives, transforms the deterministic routing problem in the wide area network into a Markov decision process (MDP), designs the action space, state space, and reward function. Additionally, a Path and Cycle Decoupled Proximal Policy Optimization Algorithm (PCDPPO) based on deep reinforcement learning (DRL) is proposed to optimize the wide-area network routing problem. Simulation results demonstrate that compared with existing methods, the Integrated Scheduling Efficiency (ISE) of this algorithm is increased by more than 4.87%, and the convergence time is less than 30.83% of the non decoupled method. This research offers an effective solution for the efficient, stable, and reliable network transmission of specific services in new power systems.

In the high-density data transmission and multi-terminal device environment of 6G edge networks, an efficient routing strategy is crucial. Existing routing methods lack adaptability in the face of network dynamics, which can result in service delays and connection interruptions. To address this issue, this paper proposes an innovative fault-tolerant traffic routing (FTTR) mechanism. Leveraging Message Passing Neural Networks (MPNNs) and Deep Q-Network (DQN), FTTR can deeply explore the interdependencies between links and make precise routing decisions. Extensive experiments demonstrate that FTTR mechanism achieves an average 27.72% increase in network load capacity compared to mainstream routing strategies. Its generalization and robustness significantly outperform Proximal Policy Optimization (PPO), clearly demonstrating its adaptability and stability to network dynamics.

When terrestrial communication base stations are destroyed, communication efficiency and emergency response capability are severely degraded. To address this challenge, we propose a rapid UAV enabled communication service within a low altitude intelligent networking architecture. The method adopts the Soft Actor Critic reinforcement learning framework and integrates a Transformer as the spatial structure feature extractor to enable intelligent UAV deployment in complex post disaster environments. Experiments on real geographic coordinate data compare Graph Convolutional Network and Graph Attention Network feature extractors with respect to coverage rate, normalized delay, and energy cost. The SAC-Transformer achieves an average coverage of 99.8%, improving over SAC-GAT, SAC, and SAC-GCN by 5.83%, 14.45%, and 15.64%, respectively. Its average delay is reduced by 39.43%, 47.80%, and 32.56% relative to SAC-GAT, SAC, and SAC-GCN, respectively. Overall, the SAC-Transformer markedly enhances coverage and robustness in complex post disaster scenarios and exhibits stronger multi objective performance with high coverage, low delay, and controllable energy cost, making it well suited for UAV emergency deployments in spatially complex settings with stringent communication constraints.

The rapid development of autonomous vehicles and smart city has led to an exponential increase in data generation within Intelligent Transportation Systems (ITS). However, comprehensive extraction and utilization of these data are severely hindered by communication and energy constraints, security and privacy concerns, vehicle mobility limitations, and spatial distribution challenges. Using 6G and Digital Twin (DT) technologies offers a promising solution to these problems. In this paper, we propose a DT-based model training architecture for vehicular networks and introduce Federated Learning (FL) to preserve data privacy. While distributed model training and parameter transmission introduce challenges in delay and energy consumption, which conflict with real-time service requirements in ITS. In addition, the quality of the data and the processing capability of each vehicle varies widely, which will affect the efficiency of data sharing and model accuracy. Therefore, it is vital to select appropriate training nodes and optimize resource allocation under the constraints of task delay and energy consumption. We formulate an optimization model to improve the selection of FL participating nodes and energy management strategies, aiming to maximize accuracy while minimizing energy consumption. We then develop a DT-assisted deep reinforcement learning (DRL) method. Experiments show that our scheme achieves higher training accuracy and energy efficiency compared to the benchmark.

To support the escalating traffic demands in the 6G era, the novel computing paradigm of in-network computing (INC), where tasks can be processed on the forwarding path, is emerging with enhanced network performance and improved service quality. Considering the large network scale with high dynamics, effective task scheduling in INC paradigm becomes imperative but challenging. In this work, we investigate the task scheduling in INC paradigm to minimize both the task delay and network energy consumption, while considering constraints on task latency, traffic dynamics, and available communication and computing resources. We first construct a novel computing and communication model considering the traffic variation in network nodes on the transmission path. To solve the task scheduling problem, we propose an algorithm, named as NBFNDRL, which is a deep reinforcement learning (DRL) algorithm based on Neural Bellman Ford networks (NBFNet). NBFNet can learn high-dimensional correlated graph structural information, utilize message-passing mechanisms to represent changes in traffic between adjacent nodes, predict scheduling paths, and provide a basis for DRL decision-making. The DRL agent trains and updates NBFNet through interaction with the environment. Finally, we present simulation results to demonstrate the effectiveness of our proposed approach in comparison to benchmark algorithms and various computing paradigms.

Intelligent technology is driving the communication industry to a higher stage of autonomy. To enable the large-scale deployment of Autonomous Networks (ANs), We design a knowledge-driven self-healing dual-loop architecture throughout the fault lifecycle, Then we propose an optimization model that aims to minimize the loss of services operation. Simulations are conducted in a programmable network to validate the loop.

Network traffic prediction is crucial for network operation and management, forming the basis for utilizing big data in decision support. Traditional deep learning methods require extensive, assumed independently and identically distributed(IID) data. However, with IoT development, privacy protection gains importance, resulting in distributed data collection with varying distributions. This leads to a significant performance drop when applying a well-trained model to a new dataset, causing domain shift. To tackle domain shift from inconsistent data distributions and meet privacy protection needs, this paper proposes a spatial-temporal feature learning method within the federated domain generalization framework for network traffic prediction. The ultimately trained model effectively generalizes to an unseen domain, as confirmed by experimental results.

Augmented Reality (AR) applications are latency-sensitive and contain significant heterogeneous content, such as mixed static objects and interactive data. Relying solely on real-time edge caching makes it difficult to meet the latency requirements of AR, disrupting user’s immersive experience. In addition, the operator can motivate terminal caching foreground content and reduce transmission costs through device-to-device (D2D). Therefore, we proposes a cloud-edge-terminal collaborative proactive caching and differentiated delivery mechanism of heterogeneous content, which reduces service response latency and improves comprehensive revenue through efficient edge collaboration methods, accurate heterogeneous content pre-caching strategies, and differentiated delivery mechanisms. Firstly, we synthetically considers AR user’s service response latency and operator’s comprehensive revenue, proposing a user behavior and resource-aware edge collaborative service domain construction method to improve the collaborative service capability of edge nodes. Then, it proposes a pre-caching algorithm for heterogeneous content based on foreground/background content separation, user preference prediction, and storage space partitioning to improve cache utilization in the edge network. In particular, a D2D-assisted differentiated delivery strategy is designed to improve service response speed and overall revenue. The numerical results show that the proposed mechanisms are better than other solutions and can improve cache hit rates and operator’s comprehensive revenue.

Remote direct memory access (RDMA) networks are widely deployed in clouds and data centers for low latency and high throughput. Emerging applications like artificial intelligence training and serverless computing demand RDMA networks with high connection scalability. However, increasing connections can reduce throughput, as many academic research states. Conversely, some industry reports indicate the scalability issues are situation-specific. Despite this, existing work lacks a comprehensive analysis of RDMA scalability issue triggers and causes. In this paper, we revisit triggering conditions and underlying causes of RDMA scalability issues. First, we comprehensively analyze RDMA data flows and potential resource contentions, particularly with direct cache access. Second, we conduct extensive tests across varied measurement settings and RDMA NICs (RNICs). We identify triggering conditions including frequent switching of queue pair connections, rapid request posting, intensive memory access demands and inappropriate program configurations. We also systematically uncover causes of scalability issues, mainly due to RNIC cache misses, RNIC processing unit backpressure, host last-level cache misses, and software inefficiency. Finally, we provide guidelines to mitigate the RDMA scalability issues.

With the rapid development of the Internet of Things (IoT), more and more IoT traffic is generated in the data network. Accurate perception of IoT traffic changes will facilitate traffic engineering decisions, thus ensuring the performance of IoT applications. However, current traffic prediction methods ignore the limitations of actual application environment. In this article, we propose an IoT traffic prediction method based on horizontal federated learning to predict traffic trends under the cooperation of the cloud and the edge side. In order to improve the accuracy of IoT traffic prediction, a traffic prediction model SMN3-CIFGA is proposed to predict IoT traffic based on traffic feature extraction in a limited hardware environment. In addition, in order to improve the communication efficiency in the distributed training process of the traffic prediction model, we propose a gradient compression algorithm based on dynamic threshold (GCADT). The experimental results demonstrate that compared with current methods, the average training time of the GCADT algorithm is reduced by about 6.21%, the transmission gradient size of the GCADT is reduced by about 66.71%, the average training time of the classification model SMN3 is reduced by about 40%, and the testing set prediction accuracy of SMN3-CIFGA can reach 97.61%.

In the era of 5G and beyond, different service scenarios have put forward rich and differentiated requirements for the carrier network. The emergence of flexible Ethernet technology has met the needs of high-speed transmission and flexible bandwidth configuration. However, the current FlexE transmission mechanism based on the 5Gbit/s granularity creates a massive waste of resources in multi-granularity hard isolation services. To optimize the utilization of slot resources, we propose a new FlexE calendar slot allocation mechanism based on a novel hybrid-granularity model. This mechanism encompasses traffic prediction and a calendar slot allocation method based on a FlexE hybrid-granularity model. The former adapts the slot allocation process to the fluctuations in client flows through accurate traffic prediction. The latter adopts a hybrid-granularity slot allocation algorithm based on dynamic programming to ensure a high isolation of the service transmissions and to improve the utilization of slots. A comparison with the existing schemes shows that under experiments with different periodic regularities, the proposed method can increase the slot utilization by 71.5%-77.9%, and under experiments with diverse client granularity distributions, the proposed method can increase the slot utilization by 59.2%-76.7%.

In the context of Software Defined Networking (SDN) scenarios, the deployment of multi-path routing has been trending as one of the practical approaches. It serves the two-fold objective of improving the reliability of Service Function Chains (SFCs) and reducing end-to-end delays through parallel processing; this latter being this paper’s focal point given it is one of the fundamental objectives of 6G. The literature encloses numerous publications revolving around the exploitation of Virtual Network Function (VNF) duplication and optimal placement to enable parallel processing. However, very little attention has been allocated to segmented VNFs with parallel multi-path data traffic flow routing to catalyze service completion. In reality, the application of segmented task processing is now widely used in our Internet life (e,g, real-time video on Youtube). In order to realize the ultra-low end-to-end delay of SFC, we introduce the segmented VNF processing window and implement VNF processing tasks in batches/windows with multi-path routing. Herein, a novel Parallel Multi-Path service flow Routing with processing Windows (PMPRW) scheme is proposed. The PMPRW is formulated as a Mixed Integer Linear Program (MILP), owing to the complexity of which, a Column-Generation (CG) based framework is developed to generate accurate sub-optimal solutions that achieve the same performance as the optimal solution. In order to accelerate the process and enhance the performance, we propose an extended Column Fixing (CF) strategy to help generate new columns in CG. Extensive simulations are conducted to gauge the merit of PMPRW and demonstrate its superiority (as opposed to single-path routing). PMPRW achieves desirable performance by concurrently reducing the overall end-to-end delay (e.g., 22% through parallel dual-path routing).

Time sensitive networking (TSN) has been widely adopted and applied in many fields. The scheduling problem of TSN requires that the gate control list (GCL) is calculated according to the flow information in a given topology network. Conventional flow scheduling schemes are usually based on the given routing scheme, which limits the scheduling performance. Besides, current works mostly focus on the time trigger flows (TT). However, AVB flows exist as aperiodic flows in the industrial Internet. The integrated scheduling of these two types of flows is required to improve the overall schedulability. In this paper, a problem model of joint routing and GCL scheduling is proposed. An algorithm based on Tabu search (Tabu-RG) is proposed to solve the problem with specific design of neighborhood movement policy, neighborhood selection policy, as well as diversified function. Experimental results show that compared with the solver method, the proposed algorithm can save 75% of the time cost on the premise of ensuring the solution performance.

Faultless and cost-saving route calculation plays a fundamental role in the network traffic engineering. The existing route calculation methods mostly are data-driven and lack the interpretability, which make less use of complex network information. Knowledge graph has the ability to convert complex network data into interrelated knowledge, so the interpretability can be enhanced. In this paper, we propose a novel network route calculation solution based on the knowledge graph, which utilizes the link prediction to handle the high-dimensional features of network. By improving the path-based link prediction framework NBFNet, the defects caused by lack of interpretability are made up. The simulation results show that the proposed method outperforms Bellman-Ford and TransH in terms of packet loss and delay and maintains the network load balance with various network topologies.

Network traffic classification is a key technology in network communication management, which is of great significance for building intelligent communication and so on. Due to the difficult and time-consuming process of network traffic labeling, it is difficult to obtain any labeled traffic data in some special networks. At the same time, in a real network environment, there are multiple network traffic domains, and the data distribution of each network traffic domain is different, making it extremely difficult to train a network traffic classification model that performs well on multiple traffic domains simultaneously. Therefore, this paper proposes a network traffic classification method in unsupervised scenarios, aiming to study how to learn traffic knowledge from multiple source traffic domains and achieve accurate classification of unlabeled network traffic without labeled traffic data in the target traffic domain. This paper divides three traffic domains from the data set, namely VPN, nonVPN and nonTor. And three traffic classification tasks of unsupervised multi-source domain are constructed. The accuracy of traffic classification tasks in the source traffic domain is nonTor and nonVPN, and the target traffic domain is VPN reaches 89.76%. The source traffic domain is VPN and nonTor, the accuracy of the classification task is 91.73% when the target traffic domain is nonVPN, and 90.35% when the source traffic domain is VPN and nonVPN, and the target traffic domain is nonTor. Experimental results show the effectiveness of the network traffic classification algorithm proposed in this paper.

With the advancement and widespread adoption of Artificial Intelligence (AI) and Internet of Things (IoT) technology, machine learning (ML) can be leveraged in au-tomated factories to enable intelligent robot fault detection and recognition. As an emerging distributed machine learning framework, federated learning (FL) enables collaborative model training while safeguarding the privacy of user data. However, FL encounters various challenges, including the lack of stable energy supply to maintain continuous training and the existence of malicious terminals in the system. These situations could result in issues such as free-riding attacks and robots running out of energy during the training process. Hence, this paper presents a reliability and energy balanced computation resource allocation mechanism in FL system for IoT businesses. Firstly, a training latency and model contribution based reputation evaluation model is proposed to reveal the reliability of IoT terminals. Then, a training client reliable selection and delayed admission strategy is proposed, which takes into account both the reputation and energy consumption, aiming at enabling terminals with high-quality data and low energy capacity to keep contributing to the model during the later training stage. Especially, we integrate reputation and computing capability of IoT terminals to decide the freshness level of the allocated FL global model, which can effectively defend against free-riding attacks. The simulation results verify that the proposed mechanism outperforms non-delayed/partial-delayed admission mechanisms in terms of model quality and system stability.

In the upcoming B5G/6G era, devices will generate a amount of heterogeneous data at the network edge. As a paradigm for implementing distributed and privacy-preserving machine learning (ML), Federated Learning (FL) has drawn great attention to secure data sharing in edge networks. However, FL takes too much time and communication resources to train and transmit model parameters, which is unaffordable for edge devices with limited capabilities. To achieve a trade-off between resource and efficiency, it is crucial to select appropriate training nodes. While existing works about node selection focus on the resources allocation and pay less attention to the node mobility and seamless service. In this paper, we considering mobility, computation capability, and transmission power of training nodes to minimize the FL system cost. We propose an algorithm and mechanism respectively for different scenarios of node speed. An algorithm based on Deep Reinforcement Learning (DRL) matches with stationary and low-speed training nodes. A heuristic mechanism is used for nodes with high mobility. Simulation results show that the proposed schemes select appropriate training nodes effectively, and reduce the system cost by up to 20%.

To both take the advantage of wired and wireless networks, the burgeoning mobile edge computing (MEC) technology is integrated into fiber-wireless (FiWi) network to support the cost-effective deployment of Internet of Things (IoT). However, the trusted model training, efficient task computing, reasonable comprehensive energy consumption and different quality of services, are still the key problems to be solved. Thus, we introduce the federated reinforcement learning (FRL) to the framework to jointly optimize the accessing mode selection, computation offloading decision and transmission power allocation without the leakage of users’ privacy. Then, we further design a twolayer FRL algorithm based on reputation value to respectively realize the protection of user privacy and efficient optimization of the global model. The simulation results demonstrate that our proposed method outperforms others in balancing energy consumption, reducing service delay, as well as providing differentiated services.

To provide intelligent and accurate assistant driving services in smart city, as well as ensure the security and tamper proof of vehicle data, this paper build a deep reinforcement learning (DRL) and main-side blockchain empowered service framework. By storing driving data and vehicle information on the sidechain, while deploying index information on the mainchain, the main-side blockchain structure can enhance the scalability of blockchain, decrease the communication overhead, improve consensus efficiency, and avoid the leakage of data between different sidechains. However, the resource limited vehicles on sidechain cannot process numerous computation-intensive mining tasks in time, resulting in high service delays. Thus, this paper integrate mobile edge computing with blockchain system to design a double-layer mining service offloading mechanism, allowing the edge nodes and neighboring vehicles to form a cooperative mining network and collaboratively participate in mining process with specific offloading rates. The first layer uses Asynchronous Advantage Actor-critic (A3C) algorithm to efficiently offload partial mining task from the task vehicle to the road side unit (RSU), and the second layer applies double auction to specifically obtain the offloading rates from RSU to multiple service vehicles. Simulation results demonstrate that, our proposed mechanism outperforms other compared algorithms in the average profit and consensus delay.

The high-precision clock synchronization is pursued with the consideration of the balance for time-domain resource utilization, when 5G technologies are expected to carry the timing services. Firstly, a clock synchronization model is established in the case that the system exists the observed value loss. The probability distribution of the loss of the observed value is evaluated by a newly proposed delay deterministic confidence method. Then, the error boundness of the clock synchronization is investigated by the Kalman filter algorithm, and the optimization problem for 5G time-domain resource configuration is formulated with guaranteeing the precision of clock synchronization; Finally, the proposed optimization problem is solved by dueling double deep Q-learning-based Lyapunov optimization. The experimental results verify the effectiveness and superiority of the proposed method in terms of the joint optimization of clock synchronization error covariance and throughput.

Small unmanned air vehicles (UAVs) have many advantages, including low cost and flexible deployment. And they play an important role to collect the sensing data of Internet of Things (IoT). However, limited by the load capability, it is a big challenge for them to perform long-term, large range, or far distance tasks. In order to tackle these challenges, we propose to use assembly UAVs, in which we can jointly optimize the resource management, especially, the energy resource. The system model, energy cyclic cooperation, and one of the typical applications based on assembly UAVs are introduced. The energy cooperation problems are formulated and an in-air replenishing strategy (IA-RS) is proposed. Simulations show that the performances of the proposed IA-RS outperform those of the traditional on-ground replenishing strategy (OG-RS). The working time of task UAV (UAV-T) could reduce 8.3%–19.5%, and the freshness of the collected data and the collecting efficiency of the UAV-T can be improved. We also optimize the path of the replenishing UAVs (UAV-Rs). Simulations show that the proposed reinforcement learning (RL) algorithm has the best performances and acceptable complexity. Consequently, the efficiency of the IoT data collection task is improved by the proposed assembly UAVs.

To deal with the explosion connections and data volume for emergency communication or hot spot capacity enhancement with massive Internet of Things (IoT) devices, deploying aerial base stations (AeBSs) on unmanned aerial vehicles (UAVs) to generate heterogeneous space–air–ground networks is considered to be a quite effective method. However, the flying AeBSs and back-hauling to existing heterogeneous networks (HetNets) lead to network energy consumption a key point. To ensure the energy-efficient operation of space–air–ground networks for smart IoT applications, we put forward the cluster-based HetNets energy-efficient resource allocation mechanism (CHERA). The scheme first divides the entire network into multiple independent BS clusters with the K-means++ algorithm for distributed energy efficiency (EE) optimization. Then, we propose a greedy BS sleeping strategy and a Lagrangian-dual-based optimal power allocation algorithm for the maximization of EE in each BS cluster. The EE optimization of space–air–ground IoT networks is implemented under the self-organizing network framework to make sure of the efficient and reliable operation of the network. Simulation results indicate that energy consumption is effectively decreased with the mechanism. It boosts the EE of space–air–ground networks by 23.8% compared with a baseline algorithm in which BSs are all in active mode with no power optimization. The result is expected to be useful for achieving future green space–air–ground networks IoT applications.

6G edge networks strive to offer ubiquitous intelligent services, requiring a greater emphasis on network stability and reliability. However, current networks present a low automation degree of the operation, administration and maintenance process. Consequently, active service migration away from abnormal network nodes and links, as well as automatic and transparent service recovery from sudden anomalies, become challenging tasks. These conditions underscore the urgency for an innovative service self-healing mechanism for 6G edge networks. Digital twin (DT) technology uses modeling to represent physical entities, thereby facilitating lifecycle management. However, the application of DT technology in networks is still a burgeoning field of study. In this paper, we explore the DT-driven service self-healing mechanism in 6G edge networks. Initially, we design a DT-based architecture for service self-healing. Subsequently, we construct a performance prediction mechanism leveraging graph neural networks (GNNs) to devise an efficient prediction model, which aims to accurately infer network performance and promptly detect abnormal network conditions. To maintain fine-grained service stability amidst potential network anomalies, we propose a DT-driven service redeployment mechanism enhanced by GNNs. Comprehensive experimental results reveal that our proposed mechanism can accurately predict flow-level delays and identify abnormal links and nodes. Furthermore, the DT-driven service redeployment mechanism effectively reduces service delay and enhances network load balance.

With the development of 5G/6G networks, the number of wireless users is growing exponentially, and the application scenarios are increasingly diversified. Using unmanned aerial vehicles as base stations (UAV-BSs) to serve ground users has become a trend for wide area coverage and capacity enhancement for rapid access of service in 6G networks. However, as UAV-BSs have limited energy or battery storage, solutions to optimize energy efficiency while providing high-quality services are necessary. Therefore, this paper mainly concentrates on the energy-efficient deployment of coverage-aimed UAV-BSs (Co-UAV-BSs) and capacity-aimed UAV-BSs (Ca-UAV-BSs) for the coverage and capacity enhancement of ground communication under disaster areas or burst data traffic. First, Co-UAV-BSs are deployed with DQN algorithm to to get the UAV-BSs’ optimal flight paths, which mainly adopted to detect out of service users in such areas. Then the users are completely clustered based on the detection results. After that, Co-UAV-BSs and Ca-UAV-BSs are deployed hierarchically based on the user distribution and sought to optimize the energy efficiency with acceptable user services. Still, DQN algorithm and the A3C algorithm are used for obtaining all the UAV-BSs’ location deployment and users’ best connections. The simulation results show that the dynamic flying path requires less energy than the fixed path for user detecting. For the coverage and capacity enhancement, it reveals the solution we proposed could provide high-quality service for users with high energy efficiency comparing to traditional algorithms.

When an online service, such as an online Web service, encounters an abnormality, operators need to analyse many abnormal monitoring indicators on the affected machine and rapidly determine the root cause indicators, which can quickly locate specific problems. Then, the abnormality is isolated until the cascading effects caused by the anomalies are eliminated. This paper proposes a root cause indicator location algorithm named ProphetKdeRCL. First, the improved Prophet algorithm detects many abnormal time-series indicators. Then, the abnormal deviation degree algorithm based on kernel density estimation is used to measure the fluctuation of each abnormal indicator for sorting. The analyses of the delay causal factor dependence are combined with the time window. Finally, a root cause location list is generated to assist operators in quick troubleshooting operations. This paper uses public datasets to evaluate the overall effectiveness of the algorithm. The results show that compared to other algorithms, the ProphetKdeRCL algorithm has higher accuracy on evaluation indicator AC@1 and superior accuracy on the other commonly used indicators, namely, AC@2 and AC@3.

Network traffic classification is a key foundation of traffic management and network security. With the development of traffic encryption technologies and more attention given to user privacy, traditional rule-based and payload-based traffic classification methods have become less effective. To address this problem, recent studies have introduced deep learning-based methods. However, most of these studies do not consider both the flow-level and packet-level characteristics, which we believe are significant in network traffic classification. To further improve the accuracy of traffic classification, this paper proposed DM-HNN, a hybrid neural network based on dual-mode features. First, we treat the packet length sequence as the flow-level feature and the initial byte of the packet as the packet-level feature. Then, we diverge into two paths to analyze the dual-mode features using neural networks. Finally, we combine the two-path features and output the final classification results. We have performed the experiments on public datasets, the results comparing to single-mode and dual-mode traffic classifiers indicate that DM-HNN can achieve excellent performance and has certain effectiveness.

5G and beyond network will support vertical industry applications, and the resource requirements of each service vary widely. The introduction of network slices provides great flexibility to the network, which can realize the differentiated customization requirements of service. However, while determining how to intelligently orchestrate the network slices is an important challenge, current solutions rarely treat multiple customized requirements of delay, bandwidth, load balancing, and slice isolation. In this article, network slice orchestration is considered from the perspective of slice isolation and cloud-edge collaboration. First, differentiated isolation level requirements are restricted to constraints, the customized isolation is realized. Second, bandwidth is saved and network latency is reduced via the collaboration of cloud and edge data centers. In addition, exclusive orchestration optimization objectives that match various service needs are proposed to distinguish the specific requirements of different slices. Finally, two deep reinforcement learning-based algorithms are proposed. The experimental results demonstrate that the proposed algorithms can optimize the objectives while ensuring differentiated isolation levels. For typical slices, the proposed algorithms respectively reduce bandwidth consumption by about 29% and 64%, reduce slice delay by about 14% and 70%, and optimize load balancing by about 17% and 23%.

Network delay is a crucial indicator for realizing delay-sensitive task offloading, network management, and optimization in B5G/6G edge computing networks. However, the delay prediction for edge networks becomes complicated due to diverse access strategies and heterogeneous services’ storage, computing, and communication resource requirements. Current GNN-based delay prediction models such as RouteNet and PLNet lack the ability to express the complex associations between links and paths, so the predicted delay is not accurate. In this paper, we propose a novel end-to-end delay prediction model named MixerNet for edge computing, which is based on the mixed multi-layer perceptron (MLP). In this model, a mixed MLP architecture is applied to represent the association between links in the network topology and various paths. Observing that each link may have different effects on various paths, a weight matrix is then defined and multiplied by the path matrix to express it. Thus, a complete mapping frame from network characteristics (e.g., traffic intensity and routing schemes) to delay indicator is constructed. Finally, we perform extensive experiments on NSFNET and GEANT2 datasets and regard RouteNet as the baseline model. Experimental results show that MixerNet can accurately predict end-to-end delay results on various network topologies and the mean absolute error is merely about 0.36%. MixerNet also outperforms the baseline model in most evaluation indicators, especially the mean square error has a 3-fold decrease in NSFNET.

The demand for satellite relay service is increasing, while the satellite network resources are limited and unevenly distributed, which pose a great challenge to task scheduling of tracking and data relay satellites. Most existing relay scheduling models are based on static antenna setup time, which has limitations in practical applications and leads to ineffective utilization of satellite resources. This paper models the task scheduling problem based on dynamic antenna setup time and splittable tasks, which maximizes the total scheduled task number and minimizes the total antenna setup time. We also propose a two-stage insertion heuristic to solve the problem. The experimental results show that the proposed algorithm can significantly improve the total scheduled task number, total antenna setup time and effective time window utilization compared with traditional methods.

Fine-grained service offloading in collaborative edge computing can make full use of the limited resource of edge nodes to achieve efficient parallel computing. It is imperative to select appropriate edge nodes for the subtask offloading in order to ensure the network’s load balance. However, there is a lack of research on computing offloading of end-to-end fine-grained services, and existing node selection algorithms can only be used in small-scale scenarios or networks with a fixed number of nodes. In this paper, we construct an end-to-end fine-grained computing offloading model, with load balancing as the optimization goal. Especially, a deep graph matching method, based on graph neural networks, is used for offloading node selection. It can be applied to dynamic and large-scale scenarios with strong generalization capability and fast execution speed. Compared with baseline algorithms, it greatly reduces the network load imbalance degree while ensuring a high acceptance ratio of services and meeting delay, location and resource constraints.

Knowledge graph completion (KGC) aims to perform link prediction to fill lost relations between entities by knowledge graph embedding (KGE). Translating embedding, as an efficient embedding method in KGE, is widely applied in numerous recent KGC models. However, these translating models may lack the ability to express various relation patterns and mapping properties for knowledge graphs (KGs). In this paper, a simple and well-performed translating model named TransC is proposed to express different relations. A multi-channel mechanism is defined firstly to constrain translating embeddings. Then a relation-aware transfer function is designed to break the expressive restriction and map triplets involving the same relation into a corresponding plane. We also mathematically prove that TransC is capable of expressing four popular relation patterns and all mapping properties. Finally, experimental results illustrate that TransC can efficiently represent the different relation patterns and properties and achieve better performance than state-of-the-art translating models.

In the industrial scenario, the requirements for network quality and data security of industrial Time Sensitive Network (TSN) services are becoming more and more stringent. 5G technology has the characteristics of wide bandwidth, low latency, and massive connections. Mobile Edge Computing (MEC) can improve the utilization of resources and the security of data, while future reducing the transmission delay. However, for industrial TSN services, a new network deployment architecture is still needed to provide reliable, flexible, and secure services for the industrial Internet. This paper proposes a local deployment architecture and adopts a deployment optimization scheme of Distributed Unit/ Centralized Unit/ User Plane Function/ TSN Translator (DU/CU/UPF/TT) convergence integration, highly integrated network elements, build DU/CU/UPF/TT compact integrated equipment, realize the idea of the separation of control plane and user plane, and realize efficient data forwarding. Meanwhile, some technologies such as Multiple-transmission/Reception Point (Multi-TRP) and New Radio Dual Connectivity (NR-DC) are used to meet the requirements of the low latency of industrial TSN services, and ensure data security and improve network quality.

With the development of smart traffic and assisted driving, the mobile edge computing and artificial intelligence technologies are seen as the key solutions in the internet of vehicles. However, the limited edge network resources and leakage of vehicle private data in assisted driving process are still problems to be solved. Therefore, we design a federated learning (FL) empowered edge collaborative content caching mechanism to provide low latency and high reliable assisted driving services for vehicles. First, we build an edge collaborative cache domain to allow multiple edge nodes to jointly share the service component resources required by vehicles. Next, based on LSTM prediction model obtained by FL, we propose a service component pre-caching and placement strategy according to the predicted and real-time vehicle behavior, to realize fast and accurate content caching services. The simulation results show that the proposed mechanism can improve the performance in terms of caching hit rate, service delay and the resource utilization of edge nodes.

Vehicle-to-everything communication is an indispensable component of 6G networks that could help to facilitate future transportation systems. However, massive vehicles and unstable vehicle-to-vehicle (V2V) links may become bottlenecks for the low-latency delivery of contents, such as safety-critical emergency messages and multimedia. Instead of resolving the problem in a centralized way, we propose a massive vehicular Internet-of-Things system and investigate the approach that would enable each vehicle to decide the transmission mode from three modes, i.e., vehicle-to-network, vehicle-to-infrastructure and V2V sidelinks, and wireless resources. Specifically, a multiagent deep reinforcement learning (RL) framework is formulated by combining the multiagent RL approach, WoLF-PHC, with the techniques from deep$Q$-learning (DQN) to gain the formulated framework with the capability of capturing the effects of interaction between learning agents and states of complex environment. The framework is set to maximize the throughput of vehicles while maintaining the latency and reliability constraints of the vehicle communication links. However, it could be easily extended to other objectives. The simulation results demonstrate that the proposed approach outperforms the compared ones in total traffic capacity and satisfaction rate of the vehicles in communication.

As an emerging distributed machine learning (ML) method, federated learning (FL) can protect data privacy through collaborative learning of artificial intelligence (AI) models across a large number of devices. However, inefficiency and vulnerability to poisoning attacks have slowed FL performance. Therefore, a blockchain-based asynchronous federated learning (BAFL) framework is proposed to ensure the security and efficiency required by FL. The blockchain ensures that the model data cannot be tampered with while asynchronous learning speeds up global aggregation. A novel entropy weight method is used to evaluate the participating rank and proportion of the local model trained in BAFL of the devices. The energy consumption and local model update efficiency are balanced by adjusting the local training and communication delay and optimizing the block generation rate. The extensive evaluation results show that the proposed BAFL framework has higher efficiency and higher performance for preventing poisoning attacks than other distributed ML methods.

The Industrial Internet of Things (IIoT) is one of the important applications under the 5G massive machine type of communication (mMTC) scenario. To ensure the high reliability of IIoT services, it is necessary to apply an efficient resource allocation method under the dynamic and complex environment. In view of the absence of energy-efficient resource management architecture for the entire network, this article proposes an intelligent-driven green resource allocation mechanism for the IIoT under 5G heterogeneous networks. First, an intelligent end-to-end self-organizing resource allocation framework for IIoT service is given. Next, an energy-efficient resource allocation model within the framework is proposed. It is then solved by an intelligent mechanism with the asynchronous advantage actor critic driven deep reinforcement learning algorithm. Through the comparison analysis of different methods and rewards under IIoT scenarios with proper parameters setting, the proposed method can achieve better performance than other traditional deep learning (DL) methods and maintain service quality above accepted levels as well.

With the emergence of industrial Internet of Things (IIoT), intensive computation workload will be imposed to industrial end units (IEUs). By leveraging mobile edge computing (MEC), the local computational tasks can be offloaded to servers deployed in mobile edge networks with low latency. This article proposes the intelligent cost-and-energy-effective task offloading in the 5G and Wi-Fi 6 coexisting heterogeneous IIoT networks. The novel joint task scheduling and resource allocation approach comprises the following two parts: a Lyapunov optimization-based component to decide local task scheduling and computing power and an online multiagent reinforcement learning component together with a game theory-based algorithm to select offloading link and decide transmit power, respectively. Simulation results demonstrate the proposed approach holds obvious advantage over the compared “intuition” and “cost optimal” approaches in the efficiency of making comprehensive decision that improves energy efficiency and cost while controlling task delay in the multi-IEU and multiaccess-node MEC systems.

In order to realize the reliable and safe operation of the smart railways, and provide high quality information transmission service for passengers, the railway system needs to develop innovative communication network and advanced communication technology to meet the gradually increasing service demand of multi-dimensional comprehensive information resources. The Space-Air-Ground Integrated Network (SAGIN) can provide seamless information services for land, sea, air and space users, and is an effective solution to the challenge posed by the future smart railways to the all-time, all-domain, all-air, high-reliability and high-throughput communication. This paper aims to comprehensively discuss the technical development and application examples of High-Speed Railways (HSRs) based onSAGIN. Firstly, we analysis the development of theSAGINand the mobile communication network of theHSRs, and comprehensively discuss the single network architecture of the space-based, air-based and ground-based networks, as well as the integrated network, and discuss the application scenario and network structure of the combination of the integrated networks. At the same time, the communication services, existing problems and key technologies of the space-based, air-based and ground-based networks are discussed, and the application trend of theSAGINinHSRsis presented. Furthermore, the application scenarios of Artificial Intelligence (AI) technologies in solving the efficient resource utilization of smart railways communication and theSAGINare studied. Based on these technologies, we point out the research direction for the future development of AI technologies inSAGINinHSRscommunications.

Energy consumption is becoming a key issue in the research of future network. In practice, network traffic has a periodic time distribution that occurs most often at a low level. This feature provides the possibility of achieving network energy savings through topology switching. By considering the deficiencies in existing studies, such as the low adaptability between network working topology and traffic load, the abnormal topology switching caused by abnormal and unbalanced traffic, and the low reliability of energy-saving topology, this paper proposes an energy-efficient routing method for software-defined networks based on topology switching and reliability. The method involves two parts: a topology-switching method and a failure recovery method. The former adapts the network working topology to the network traffic demands through dynamic topology switching to decrease the network energy consumption. The latter adopts an active strategy for fast fault recovery to ensure network reliability in the energy-efficient topology. Two network typologies and their traffic data are used to experimentally verify the method. The results show that, compared with the static topology switching method TLS, the energy saving of the proposed method can be improved at most 2.07 times and 4.63 times in two typical typologies, respectively, while ensuring network reliability.

5G realizes flexible networking by building network slices, and its realization depends on network function virtualization (NFV) technology, which combines different types of virtual network functions (VNFs) to provide network services. The reliability of VNFs is lower than that of traditional hardware due to the risk of both software and hardware failure, and redundant backup is an effective solution. Meanwhile, from the security point of view, because the 5G network is based on the unified and standardized hardware of the industry, the need for isolation is put forward. Current research on VNF reliability assurance has not considered the special isolation requirements of 5G. In this paper, aiming to guarantee the safety demand as well as minimize backup resource to meet the reliability target, we formalize the safety-oriented backup problem for 5G core network slices and propose a backup algorithm based on isolation (BABI). Simulation results show that the introduction of isolation can double the security of slices. The comparison with the existing backup methods shows that under the same isolation constraint, the proposed approach can achieve a less resource consumption by 60% - 80% and a improvement of the proportion of effective resources by 40% - 80%.

With the development of 5G, the distribution of base stations tends to be dense. Compared with the traditional network architecture, Cloud Radio Access Networks(C-RAN) architecture can satisfy the current requirements of high bandwidth, low latency and low energy consumption. Currently most energy-saving scheme for C-RAN is complex with time cost computing, which may not be suitable for large-scale region. For the problem of energy-efficient resource allocation for dense distribution of Remote Radio Heads(RRHs) in C-RAN, we use K-means clustering algorithm to simplify the network topology and reduce the complexity under a distributed manner. Aiming at the problem of network resource allocation in C-RAN, we use A3C algorithm to allocate network transmission power, and compare the total energy consumption, system energy efficiency and Signal to Interference plus Noise Ratio(SINR) value of terminal devices through simulation experiments. The experimental results show that in the same network environment, A3C algorithm has the highest energy efficiency, and can keep the SINR value of terminal devices in a reasonable range, which proves the effectiveness of A3C algorithm.

Deploying aerial base stations (AeBSs) has been regarded as an effective solution to wireless network capacity enhancement in specific areas with excessive traffic burden but insufficient capacity. Since the traffic distributions in wireless networks tend to be ever-changing, the deployed AeBS need to adjust its position to rapidly and continuously adapt to the drifting capacity enhancement demands, which is difficult to handle with traditional optimization methods due to high computational complexity and poor scalability. In this paper, we design a multi-agent deep reinforcement learning-based 3D AeBS deployment algorithm with the goal of maximizing the system throughput, which is able to make decisions in dynamic environments and conducts in a distributed manner. Additionally, in order to address the interference issue between multiple AeBSs, we adopt the Coordinated Multiple Points Transmission (CoMP) in the air-to-ground communication and propose a clustering algorithm to form groups of AeBSs for cooperative communication based on the network interference characteristics. Simulation results demonstrate that the proposed approach has significant throughput gains over conventional schemes without CoMP, and that the proposed multi-agent deep Q network (MADQN) is more efficient than centralized DQN in deriving the solution.

With the continuous construction and development of smart grid, the continuous introduction of new services, largescale access of new energy sources, energy storage and charging piles, and continuous growth of large-bandwidth services, traditional communication networks have been unable to meet the requirements of smart grid. FlexE technology adds FlexShim layer in MAC layer and PCS layer to achieve network flexibility, sub-rate, rigid interface and other characteristics, which can be well connected with IP/Ethernet technology to meet the higher requirements of network bandwidth, delay, slicing, reliability and other aspects. Therefore, it is necessary to develop a slicing network framework based on flexible Ethernet technology oriented to the communication requirements and trends of smart grid. Based on the analysis of the differentiated (deterministic time delay, large bandwidth, security isolation, etc.) requirements analysis of power communication services, we proposed a slice network framework based on flexible Ethernet technology and studied typical use cases based on FlexE technology for power services.

With the advent of the 5G era and the smart grid era, in order to achieve high reliability of grid power supply efficiency, the combination of power Internet of Things with artificial intelligence, edge computing, and advanced communication technologies is the basis for the interconnection of everything in the smart grid era. As an important tool for edge computing-oriented perceptual access implementation, edge IoT agents can not only have gateway functions such as protocol conversion and data collection, but also carry applications including edge-side data storage and stream data processing, intelligent reasoning decision-making, etc. service. Traditional edge IoT agents mostly use heavyweight virtual machines as the implementation technology, and the applications provided are tightly coupled, and they cannot achieve mutual isolation and independent deployment between applications. Therefore, this paper uses lightweight virtualized Docker container technology to deploy services and build an edge IoT agent platform based on Docker containers. At the same time, facing real-time changing access requirements, edge IoT agent clusters may have the problem of limited container load. We propose a dynamic container scheduling method to improve the access carrying capacity of container clusters and ensure the high availability of edge IoT agents.

With the rapid development of 5G commercialization on a large scale and edge computing, cloud-edge collaboration technology has been widely used in various industries, and how to achieve high efficiency cloud-edge network environment has become a research hotspot. In this paper, we propose a network efficiency evaluation model based on analytic hierarchy process (AHP) and logistic regression (LR) algorithm in cloud-edge collaborative environment. Using AHP to calculate the weight matrix of indicators and formulating multiple discrete parameter measures into the same dimensional area to obtain the final comprehensive efficiency value of the network. It describes the efficiency value in cloud-edge collaborative environment as a qualitative concept and realizes the qualitative evaluation in cloud-edge collaborative environment. On this basis, comparative experiments are carried out to evaluate the comprehensive efficiency of the network by using the traditional method and the method proposed in this paper, which verifies the practicability and effectiveness of the efficiency evaluation method.

Aiming at the problem that existing association rule mining algorithms cannot quickly mine faulty association rules in the current perception layer of electric power sensor networks, an improved eclat mining algorithm fast_eclat is proposed. The algorithm combines the characteristics of sparse data and large number of transactions at the perception layer of the power sensor network, and adopts a set intersection strategy based on pruning cross-counting, which reduces the computational complexity and improves the computational efficiency of the algorithm, which can more effectively deal with fault association rules. Comparative analysis through simulation experiments shows that the fast_eclat algorithm has better performance in the face of sparse data and large number of transactions.

In the network function virtualization (NFV) environment, network services are usually provided in the form of service function chains (SFCs), which defines the link order of virtual network functions required in service requests and are mapped to the physical network. Although NFV facilitates the flexible provision of network services, service interruptions may occur as a result of software and hardware failures. Current solutions mostly use the backup method to ensure the reliability of SFCs. However, these methods ignore the SFC construction phase that has an impact on reliability. Besides, the resource efficiency still requires improvement. To address these issues, reliability-oriented SFC construction and backup problems are investigated in this work. First, an instance-sharing and reliable construction algorithm (ISRCA) is proposed to aggregate multiple SFCs into a service function graph (SFG), and perform reliability screening for the SFG set. After mapping the SFG to the physical network, a node-ranking algorithm with centrality and reliability (NRCR) is proposed for backup node selection and backup instance deployment to improve the reliability of SFCs that have not met the requirements. Experimental results demonstrate that under the premise of ensuring reliability, the proposed backup method can reduce the consumption of bandwidth resources by about 11.7%, when combined with the proposed construction method, it can further reduce the backup resources by 13.9%.

The delay tolerant networks (DTN), which have special features, differ from the traditional networks and always encounter frequent disruptions in the process of transmission. In order to transmit data in DTN, lots of routing algorithms have been proposed, like “Minimum Expected Delay,” “Earliest Delivery,” and “Epidemic,” but all the above algorithms have not taken into account the buffer management and memory usage. With the development of intelligent algorithms, Deep Reinforcement Learning (DRL) algorithm can better adapt to the above network transmission. In this paper, we firstly build optimal models based on different scenarios so as to jointly consider the behaviors and the buffer of the communication nodes, aiming to ameliorate the process of the data transmission; then, we applied the Deep Q‐learning Network (DQN) and Advantage Actor‐Critic (A3C) approaches in different scenarios, intending to obtain end‐to‐end optimal paths of services and improve the transmission performance. In the end, we compared algorithms over different parameters and find that the models build in different scenarios can achieve 30% end‐to‐end delay decline and 80% throughput improvement, which show that our algorithms applied in are effective and the results are reliable.

With the rapid development of data centers, the energy consumption brought by more and more data centers cannot be underestimated. How to intelligently manage software‐defined data center networks to reduce network energy consumption and improve network performance is becoming an important research subject. In this paper, for the flows with deadline requirements, we study how to design the rate‐variable flow scheduling scheme to realize energy‐saving and minimize the mean completion time (MCT) of flows based on meeting the deadline requirement. The flow scheduling optimization problem can be modeled as a Markov decision process (MDP). To cope with a large solution space, we design a DDPG‐EEFS algorithm to find the optimal scheduling scheme for flows. The simulation result reveals that the DDPG‐EEFS algorithm only trains part of the states and gets a good energy‐saving effect and network performance. When the traffic intensity is small, the transmission time performance can be improved by sacrificing a little energy efficiency.

Traffic engineering (TE) can balance the traffic in the network to reduce network congestion and improve network resource utilization. The emergence of Software Defined Network (SDN) provides a more flexible and effective way to control traffic in the network. Existing TE solutions mainly focus on routing traffic via the shortest path or evenly distributing the traffic among multiple available paths, but these methods are not flexible since this static mapping of traffic to paths does not consider either the current network utilization or traffic load. Heuristics-based TE methods depends on operators' understanding of the workload and environment. Designing and implementing those methods thus take at least weeks. Furthermore, it usually takes minutes to output the solution. Inspired by recent successes in applying Deep Reinforcement Learning (DRL) techniques to solve complex control problems, we leverage DRL to control traffic in SDN. We start by building a framework which integrates the DRL algorithm into SDN. Based on this framework, we propose a modified DRL algorithm to control the traffic split ratio to multiple paths. Simulation results show that the proposed approach performs better than three baseline methods when the traffic load is dynamically changing.

To achieve reliable service transmission in such scenarios such as satellites and radio stations, Delay Tolerant Network (DTN) can be used as an effective processing method. However, the existing researches lack consideration of service-oriented end-to-end transmission guarantee method, so ensuring the reliability and continuity of DTN end-to-end service transmission has become an important issue to be solved. This paper first analyzes the characteristics and application scenarios of DTN, and then establishes the optimization models in different scenarios with the goal of maximizing service transmission throughput. And then implements dynamic routing based on node behavior analysis (DRNA), which determines routing paths according to different node behaviors. Finally, it is found that when the road is interrupted, DRNA algorithm can not only reduce the end-to-end delay of the service transmission, but also guarantee the service delivery rate.

Blockchain has broad application prospects in Smart City, and the technical characteristics of the blockchain itself can solve the problems of mistrust of network resource production relations and unfair distribution of revenue. However, most of the devices in Smart City are mobile devices with insufficient resources. The demand for computing power of the mining process cannot be met. For this, we introduce mobile edge computing, deploy edge servers on the edge side, and provide resources for mobile devices. We build an edge network resource allocation model for mobile blockchain to realize the effective application of blockchain technology in mobile environments. The resources required by the mining process can be obtained from neighboring resource sharing devices or edge servers. The resource allocation between adjacent devices can be modeled as a two-way auction model, and the Bayesian- Nash equilibrium is solved to determine the optimal price, while considering the trusted value of the device; the process of the mobile device acquiring resources from the edge server can be modeled as a two-stage Stackelberg game. Finally, simulation experiments show that this mechanism achieves a higher personal utility than an existing model that only considers requesting resources from an edge server.

In modern grid systems which is a typical cyber-physical System (CPS), information space and physical space are closely related. Once the communication link is interrupted, it will make a great damage to the power system. If the service path is too concentrated, the risk will be greatly increased. In order to solve this problem, this paper constructs a route planning algorithm that combines node load pressure, link load balance and service delay risk. At present, the existing intelligent algorithms are easy to fall into the local optimal value, so we chooses the deep reinforcement learning algorithm (DRL). Firstly, we build a risk assessment model. The node risk assessment index is established by using the node load pressure, and then the link risk assessment index is established by using the average service communication delay and link balance degree. The route planning problem is then solved by a route planning algorithm based on DRL. Finally, experiments are carried out in a simulation scenario of a power grid system. The results show that our method can find a lower risk path than the original Dijkstra algorithm and the Constraint-Dijkstra algorithm.

With the continuous development of IoT, smart grid receives widespread attention. Aiming at the problem of poor performance of the Rake receiver in a wideband micro-power wireless communication system under a low signal-to-noise ratio, the non-stationarity of the Chirp signal used in the system and the adaptability of the non-stationary signal of the denoising method are proposed. A complementary empirical mode decomposition (CEEMD) combined with wavelet threshold denoising algorithm to improve the receiver's signal-to-noise ratio. The CEEMD algorithm can not only handle non-stationary signals well, but also overcome the modal aliasing phenomenon. However, using only the CEEMD algorithm, some effective information will be lost when removing high-noise high-frequency IMF components. Therefore, this paper combines CEEMD decomposition with wavelet threshold denoising, and performs wavelet threshold denoising processing on high-frequency IMF components decomposed by CEEMD to extract useful information from high-frequency components. Through matlab software simulation, the signal-to-noise ratio has been improved by about 1dB.

IoT is widely used and plays an critical role in transforming traditional industries, leading emerging industries, improving people's lives, and providing national security. From the physical layer of communication, it is particularly important to ensure the reliability of transmission in the IoT. Polar code has outstanding performance, but its coding structure determines that the length of polar codes must be the power of 2. The structure of polar code is more flexible with the existing puncturing schemes, but the decoding performance of these schemes varies with the number of puncturing bits. The decoding performance of the puncturing scheme in C0 mode drops sharply when the number of puncturing bits is large, and the decoding performance of the puncturing scheme in C1 mode is poor when the number of puncturing bits is small. In this paper, an improved polar code puncturing scheme is proposed based on the forward sequential puncturing scheme and the bit reversal puncturing scheme in the C0 puncturing mode. Compared with the traditional puncturing scheme, the simulation results indicated that the proposed scheme solves the problem of decoding performance degradation when the number of puncturing bits is too large in C0 mode; compared with the scheme in C1 mode, the scheme in this paper has a decoding performance gain of about 0.2dB when the block error rate reaches 10-3, which can also be achieved in high or low code rate.

The beam coverage directivity of mmWave communication brings great challenges to the application research of high-speed Internet of things (IoT) terminal devices, such as unmanned vehicles and unmanned aerial vehicles (UAV). Most of the existing millimeter wave beam tracking algorithms aim at AOA /AOD (Angles of Arrival/Angles of Departure) for continuous tracking estimation. However, the estimation error will increase rapidly with the increase of AOA/AOD change speed. Inspired by Auxiliary Beam Pair (ABP) algorithm, a robust two-stage beam tracking algorithm is proposed in this paper. First, AOA/AOD is estimated by the Unscented Kalman filter (UKF) algorithm, and then the AOA/AOD is modified by the improved ABP algorithm. The simulation results show that when AOA/AOD changes at a speed greater than 0.25 degrees per time slot, the proposed two-step beam tracking algorithm significantly reduces the estimation error and effectively increases the robustness of the same type of algorithm.

In the face of rapidly emerging and explosion IP services, 5G IP+optical communication network architecture will become an important mode of communication for smart grid communication network. Under the control of SDN, management and maintenance of IP+optical networks can be realized effectively. In order to improve the collaborative ability and resource utilization of 5G IP+optical networks, this paper combines the characteristics of IP services. Firstly, risk equilibrium index is designed according to the bearing characteristics of IP network and optical network. Then, combined with network delay, bandwidth, website level difference and similarity of primary and alternate routes, a reasonable primary and alternate routes allocation model is designed. Finally, a co-allocation algorithm of service routing in 5G IP+optical networks based on deep reinforcement learning is proposed. The simulation results and comparative analysis show that the method not only fully utilize the resources of IP+optical networks, but also guarantee the average service delay and reduce the network risk. Otherwise, this method effectively improves the convergence speed, which provides demonstration and theoretical guidance for the construction of the future power communication network.

With the rapid development of communication technology, communication networks are playing an increasingly important role in people's lives. Effective management of increasingly complex networks can improve the efficiency and stability of network operations. Fault management is one of the important functions of network management. The analysis of the alarms generated in the network can dig into the underlying rules to provide useful information for fault management. However, the existing alarm association rule mining algorithms often have the problem of parameter rigidity. In this paper, a two-level windows based alarm transaction extracting algorithm is proposed to solve the problem of low efficiency when using fixed size windows. Then this paper proposes an experience extraction method based on alarm priority in deep Q network (DQN), which can calculate the sampling probability according to the importance of alarm when the memory unit enters the queue. Aiming at the rare item problem caused by the fixed support threshold in association rule mining, the improved DQN is used to dynamically adjust the minimum support in rule mining algorithm. Experimental results show that the algorithm proposed in this paper can effectively improve the efficiency of alarm transaction extraction and the accuracy of alarm association rules mining.

In order to resolve the issue of high-reliability communication over long distances, the latest Internet of Things (IoT) technology plays a key role. Effective channel estimation is the key to the overall system implementation. Aiming at the problem that the Least Square (LS) estimation algorithm in IoT is affected by noise and estimation accuracy is relatively poor. To solve this problem, an improved LS estimation algorithm in view of wavelet denoising is proposed. At first, the improved algorithm uses the LS algorithm to perform the initial estimation of the channel, and then shifts to the wavelet domain for threshold denoising. By improving the denoising threshold function, the noise is better eliminated and the estimation accuracy is improved. The bit error rate (BER)and the mean squared error (MSE) of the algorithm were simulated by MATLAB. The simulation consequent indicates that the performance of channel estimation algorithm in the paper is notably better than LS estimation algorithm, LS based on DFT denoising, and soft threshold wavelet denoising algorithm. And compared with the soft threshold wavelet denoising algorithm, the SNR of the improved algorithm is increased by about 2 dB for the same BER.

Smart grid deploys a large number of smart terminals and sensing devices to form an edge network, as well as a virtual network of information space and the power Internet of Things. As a key component of 5G and future network, the latency of end-to-end and the traffic of backhaul link could be reduced by edge network. Nevertheless, the function of storage and computing are moved down to the edge nodes in mobile edge network which increases the complexity of resource management. So it is an important issue to find out a more effectively resources allocation mechanism as well as meeting the requirements of each user. Edge computing refers to the processing of large amounts of edge data in the edge space in the edge network, thereby reducing dependence on the data center, achieving limited self-governance of the edge network, and reducing off-line threats. Although Deep Reinforcement Learning (DRL) has been applied to many of the work related to edge networks, there lacks the applications for green resource allocation. A Deep Reinforcement Learning (DRL) based green resource allocation mechanism is proposed in this paper which aims at efficiently allocating the resources while satisfying the needs of mobile users. The value of energy efficiency can be obtained when the algorithm achieves convergence according to the simulation results. The efficiency of the DRL-based mechanism and its effectiveness in meeting user requirements and implementing green resource allocation are validated.

Distant supervision relation extraction is an effective method to extract the real relation between entities from unstructured corpus. However, affected by the hypothesis of distant supervision mechanism, relation extraction model often faces the disturbance of mislabeled data and noise samples. In order to alleviate the above problems and improve the performance, we propose a relation extraction framework based on Bert-based pre-trained models, Bert for Relation Extraction (BRE). BRE uses BERT as feature extractor and loads pre-trained parameters for fine-tuning. It integrates external semantic knowledge with entity relation knowledge in specific tasks to improve the performance of classifier. In addition, we designed position enhanced CNN module and time-decay selective attention mechanism for BRE to bridge the semantic gap between external knowledge and relation knowledge, and alleviate the problem of mislabeling and noise in the multi-instance learning mode. We conducted experiments on NYT-10 and GIDS datasets, and the results show that BRE achieves the best performance.

The cloud computing market is dynamic, distributed, and lacks central authorization. In this environment, cloud resource providers are vulnerable to deception and cloud resources may be abused. How to implement efficient and feasible trusted negotiations with users to expand Benefits is an urgent issue. Based on SLA (Service Level Agreement), this paper proposes a trusted negotiation method to optimize cloud resource allocation from the perspective of cloud resource providers. In a nutshell, it firstly quantifies each indicator based on the total amount of cloud resources requested by the user and the corresponding price, the user's comprehensive credit, and the total amount of resources corresponding to each SLA level, then filters the users who meet the requirements. Next knapsack algorithm and the greedy algorithm based on dynamic programming are used to predict the allocation of cloud resources respectively. Finally, the allocated users are negotiated to reach a transaction. This article takes the resource allocation price, negotiated price, and negotiated success rate as the evaluation index. The simulation results show that compared with the greedy algorithm, the algorithm in this paper has higher resource allocation price, negotiated price and negotiated success rate under different numbers of users, and can effectively realize the optimal allocation of cloud resources.

Network Function Virtualization (NFV) is an important shift in telecommunication service provisioning. It enables the decoupling of network element functions and dedicated hardware devices. How to economically place and chain Virtual Network Functions (VNFs) according to the requirements of Service Functions Chains (SFCs) are the challenges for NFV orchestration. In this paper, we consider the offline deployment issue from the perspective of sharing VNF instance to improve resource utilization and reduce total placement costs. Firstly, we generalize the problem as a Facility Location Problem and propose a Mixed Integer Linear Programming (MILP) model. Besides, our model can be dynamically configured according to the different deployment preferences. Then we propose a heuristic algorithm based on the Steiner Tree Problem and Markov Decision Process (MDP). We evaluate our heuristic algorithm by comparing with the optimal solution of MILP and a classic graph based algorithm. The results show that the difference of the deployment costs between our algorithm and the optimal solution is less than 3%. However, the execution time can be significantly reduced by 57.4%.

To satisfy ultra-high data volume and traffic density transmission requirements, millimeter wave (mmWave) and device-to-device (D2D) communication technology will be widely used in 5G mobile communication networks. In scenarios where mmWave small cell and D2D transmission coexist, D2D links mostly reuse frequency resources of the small cell to obtain higher spectral efficiency. However, this will make D2D impose great interference to mmWave small cell. This paper designs a Stackelberg game based interference control scheme with full frequency reuse in the context of D2D underlaying mmWave small cell network. The scheme aims to optimize the transmit power of D2D links, alleviate the interference caused by D2D communication to the mmWave small cell and take full advantage of the bandwidth of the millimeter band. Simulation results show that the proposed scheme converges rapidly, keeps signal to interference plus noise ratio (SINR) in a high range and achieves excellent throughput performance.

With the development of modern network, the demand of users has increased dramatically, and more data and services are required. This has caused tremendous pressure on the Macro-cellular network of infrastructure. Air access has become a new solution for the development of communications. Unmanned aerial vehicle (UAV) is used as an air node to improve coverage and capacity. Based on deep Q-Network (DQN) algorithm and considering the different quality of service requirements of different users, this paper proposes an optimal 3D location planning algorithm. The results show that the use of multiple UAVs can not only provide capacity enhancement, but also meet the different QoS requirements of different users. The average spectral efficiency of the system is increased by 15.5%, and user coverage to meet QoS requirements increased by 25.5%.

As one of the key technologies of 5G, massive IoT environments provide the ubiquitous IoT services. Compared with 4G, its structure is more complex, and it has a large number of deployed nodes. If a failure occurs and can't be alleviated its effect in time, it will lead to a significant drop in network performance. Therefore, the cell outage compensation (COC) problem in massive IoT environments is very important. Although deep reinforcement learning (DRL) has been applied to many scenarios related to the self-organizing network (SON), there are fewer applications for cell outage compensation. In this paper, aiming at the cell outage scenario in massive IoT environments with the goal of maximizing the connectivity of base stations while meeting service quality demands of each compensation user, we present a framework based on DRL to solve it. Specifically, we first allocate compensation users to adjacent BSs by using the K-means clustering algorithm, then use DQN to find the antenna downtilt and the power allocated to compensation users. The simulation result shows that the algorithm converges quickly and tends to be stable, and reach 95% of the maximum target value. It verifies the efficiency of the DRL-based framework and its effectiveness in meeting user requirements and handling cell outage compensation.

Network Function Virtualization (NFV) promises a significant advantage for IoT operators to steer substantial customizable service through a sequence of virtual network function (VNF). Service Function Chain (SFC) mapping is a key problem in IoT network resource allocation. There are two challenges in virtual resource allocation include: (1) how to map SFC requests to appropriate devices in the right sequence; (2) how to assure QoS requirements of SFC requests. Therefore, to meet the sharp increase of IoT traffic amounts and the diversification of IoT service requirements, a multi-objective service function chain mapping mechanism is proposed with two sub-mechanisms. First, a SFC mapping algorithm is designed to embed VNFs onto the substrate layer based on cost and load balancing. Then a reliability-aware SFC backup algorithm combining SFC backup and VNF backup is presented to economically and efficiently improve service reliability. The simulation results show that the algorithm can significantly improve the acceptance ratio of SFC requests, reduce cost, ensure network balance, and achieve long-term sustainable operation of the network.

With the development of 5G and Internet of Things technologies, some studies consider combining fog computing with mobile ad hoc networks (MANETs) to form a cloud-assisted mobile ad hoc network. But it faces many challenges, such as terminal mobility, dynamic topology, multi-hop nature in transmission, limited bandwidth and battery. So, to better utilize the resource, a clustering algorithm based on communication overhead and link stability is proposed with two sub-stages. First, in clustering stage, we design a clustering method based on multiparameter-limited overhead to select resource directory index nodes for resource information management. Then, in the maintenance stage, we present a network clustering adaptive adjustment algorithm based on link stability. At last, the simulation result shows the proposed algorithm can reduce the communication overhead and improve the stability of the system.

As an effective way of utilizing spectrum resources, the application of cognitive radio technology in smart grid has been widely studied, especially in the wireless Neighborhood Area Networks(NAN). The traditional static allocation method does not consider the dynamic changes of the spectrum environment, and cannot effectively utilize the spectrum resources. This paper proposes a dynamic spectrum allocation method in the NAN scenario, to ensure that spectrum resources can be fully allocated and utilized under the constraints of the different requirements for quality of services (QOS). Firstly, a service arrival model is established considering the uncertainty of the arrival of the primary user (PU), so that the number of reserved channels can be dynamically determined based on this arrival probability model. Secondly, we propose a novel dynamic spectrum allocation method using spectrum reservation to improve the reliability of some services by sacrificing the real time of other services. Finally, the optimal number of leased channels can be determined by the stationary distribution in Markov model we proposed, in which we use blocking rate and dropping as the indicators. Numerical results show that the proposed dynamic spectrum allocation method is superior to comparison method, and some meaningful conclusions are drawn after the observation of our experiments and simulations.

Mobile edge computing in 5G network is emerging as a very promising computation architecture by pushing computation and storage closer to end users with both strategically deployed and opportunistic processing and storage resources. Baidu cloud provides network services which can be deployed in 5G network recently. The network services such as weather forecast service and city road map service are typical applications for smart city. We analysis Baidu website data in this paper by our data mining method and related software. Clustering, outlier detection, prediction, and statistical methods are used to evaluate these smart city services, and the analysis result give suggestions to improve design and development of our 5G services (API website).

Energy efficiency is essential for massive machine-type communication (mMTC), because of the limited energy in internet of things (IoT) devices. We consider a two-hop amplify-and-forward (AF) relay communication, and allow IoT devices with inferior channel conditions to connect relays by using device-to-device (D2D) technology. This paper proposes to jointly optimize relay selection, channel allocation and power control, so that the total energy efficiency is maximized while guaranteeing the signal to interference plus noise ratio (SINR) requirements of relays and BSs. The formulated joint optimization problem involves a nonlinear fractional programming (NFP) problem and a user-relay-channel matching problem which is NP-hard. Therefore, we propose a two-stage approach composed of the Dinkelbach method and a hypergraph-based 3D matching (HGM). Simulation results show that the total energy efficiency under the HGM is 8.41% and 59.85% higher than the iterative Hungarian method (IHM) and the minimum zero surface prioritized allocation (MZPA), respectively.

The satellite is a primary means for providing emergency communication backhaul in disaster areas, where large bandwidth is demanded to support communication services in a wide affected area. Millimeter-wave (mmWave) communication with sufficient spectral resources promises significant enhancement to satellite-terrestrial link capacity. However, the alignment delay and mutual interference caused by directional communications with narrow beams severely limit the capacity of mmWave communication. To this end, we optimize the beamwidth to reduce the impact of beam alignment overhead on capacity. Then, considering the multi-user interference between beams, we propose a transmission scheduling scheme based on beam sharing, namely users with strong mutual interference when served simultaneously by independent beams, share the same beam. A heuristic algorithm is proposed to derive the groups of users sharing beams, and their beamwidth. Simulation results show that the proposed scheme achieves considerable capacity enhancement compared to the one-to-one beam occupation scheme (OB) and fixed beam scheme (FB), thus improving the spectrum efficiency of mmWave satellite-terrestrial communication.

Wireless-based neighborhood area network (NAN) plays an increasingly important role in smart grid (SG) since the rapidly emerging smart services and rising number of terminals in grid put forward higher demand for NAN. Considering the differential business demands in NAN, we focus on the wirelessly uplink resource allocation, which allows a trade-off between network throughput and service fairness. For more flexible and coordinated allocation, this paper introduces the cloud-radio access network infrastructure into NAN with orthogonal frequency division multiplexing passive optical network (OFDM-PON) as the fronthaul link, and proposes a corresponding uplink resource allocation method that balances the network throughput and allocation fairness. By utilizing a hybrid intelligent optimization algorithm, composed by adaptive genetic algorithm and binary particle swarm optimization, the optimal throughput-fairness trade-off solution can be obtained with a good convergence ability. Simulation results demonstrate the advantages of our proposed method in both improving network throughput and achieving the trade-off between throughput and fairness.

In cyber-physical power system (CPPS), the communication transmission links carry key services. However, some existing routing approaches cause service routes to be too centralized in the links, which leads to increase the risk of service and network. In order to reduce the risk impact of communication transmission links interruption on the service, this paper proposes a service routing reconstruction approach based on risk balance. Firstly, we analyze the risk transmission of cross-space in CPPS. Then this paper establishes a risk assessment model by characterizing the node risks with the station load pressure, expressing the link risks with the service average communication delay and the service risk balance degree. Further, the improved genetic algorithm is adopted to solve the service routing reconstruction. Finally, based on part of power grid topology from a Chinese province, the simulation results show that the proposed approach can find a route allocation scheme with lower risk value than the original Dijkstra algorithm and genetic algorithm using "roulette wheel" selection strategy, as well as ensuring the stable operation of the power system.

System Protection Communication Network (SPCN) is a new type of high-speed, real-time, secure and reliable communication network proposed in China supporting services such as AC/DC control, pumped storage control etc. In order to reduce the impact of SPCN failure on electric power system, this paper proposes a risk modeling and optimization approach. Firstly, we build a risk model to analyze the dynamic link and service risk from aspects of failure probability and its impact value. Then, we construct a risk optimization problem aiming at minimizing the link risk balance degree with service quality and risk constraints, and propose improved genetic algorithm to solve it. Based on part of network topology from a Chinese province, simulation results show that the proposed approach can make SPCN more reliable comparing to other methods when link failure occurs.

The all-terminal reliability is the probability that all nodes in the whole network remain connected, and it is instructive to analyze the operational risk of the whole network. The exact calculation of the all-terminal reliability is an NP-hard problem, and it's not suitable for analyzing the reliability of the large-scale network. This paper presents a method of calculating the all-terminal reliability of the large-scale smart grid communication system (SGCS) based on the complex network theory. The network is divided into many communities before calculating the all-terminal reliability, and the all-terminal reliability of communities is calculated instead of the whole network. The use of complex network community structure of the SGCS can effectively reduce the complexity of the topology, and simplify the complexity of the all-terminal reliability's algorithm. Compared with the traditional approximate all-terminal reliability calculation methods, the algorithm has greatly reduced the time complexity and improved the accuracy of result.

The existing network risk control mechanisms are lack of scientific and normative decision-making and rely too much on subjective judgments, which brings a great uncertainty on network risk management. In this paper, a risk control decision-making mechanism of power data network based on grey relation is put forward, and puts emphasis on the prior risk control based on the prediction results. This mechanism first constructs a matrix of positive and negative ideal measures according to the risk control objective. Then, the grey relation coefficient matrix between the candidate and ideal measures is calculated to evaluate the similarity between measures. Finally, we define the grey relation projection coefficient to evaluate the degree of closeness between the candidate measure and the positive ideal measure and the degree of deviation between the candidate measure and the negative ideal measure. Simulation results show that this mechanism can make timely and accurate decision-making of network risk control measures.

Recently, a new opportunity for on-grid energy saving is enabled by the green network infrastructure sharing. This paper mainly investigates the collaboration between multiple operators to improve the energy utilization in this scenario. Then, an energy-saving management mechanism is proposed to reduce energy consumption and optimize energy utilization. We decompose the problem into two sub problems for base station sleeping and green energy allocation. And the BS sleeping algorithm and the green energy centralized allocation algorithm are respectively proposed to solve them. Comparing with other mechanisms, simulation results show that the proposed energy-saving management mechanism can effectively reduce 65% on-grid energy consumption while guaranteeing the quality of service (QoS) to the user equipment device (UE).

With the proliferation of bandwidth-demanding mobile applications in the era of 5G, the aggregation of a few users may lead to extremely high load in cellular base stations, producing traffic hotspot in wireless networks. Therefore the higher requirement is imposed on the flexibility of a 5G network, namely the capability of performing rapid capacity enhancement in hotspot area, which makes hotspot localization and critical prediction functions. In this paper, we proposed to localize hotspots with Gaussian Random Field (GRF)-based spatial traffic density model deduced from load data of base stations, together with the prediction with Holt-Winters. We measured the spatial traffic in a specific area within a short time span and forecasted the spatial traffic density distribution. Numeric results show the proposed approach can localize hotspot efficiently, and during traffic peak hours, hotspot prediction is of high success rate.

Cognitive-radio-based smart grid networks have been studied recently as an efficient way to overcome radio spectrum shortages, especially in wireless Neighborhood Area Network (NAN). In this paper, we propose the optimal spectrum allocation strategy of cognitive radio NAN Gateway (NGW), which also acts as a spectrum collector by radio sensing and leasing from the providers for a fee. Since the service terminals in grid are heterogeneous based on different QoS requirements and willingness to pay, this paper uses differential pricing and admission control for different terminals to improve the benefits of NGW. The decision-making process for spectrum collection and allocation is modeled as a 4-stage Stackelberg gaming, where the optimal decision of radio sensing, spectrum leasing, admission control and differential pricing are deducted through a reverse derivation. A novel corresponding algorithm is also given to solve these optimal solutions efficiently. The numerical results verify the theoretical work sufficiently, meanwhile some obvious meaningful conclusions are drawn from the observation of numerical experiments.

This paper studies the multi-layer structure of coupled power network based on the problem of fault cascade and unreasonable network design in the network virtualization environment (NV). Based on the complex network theory, we propose a network optimization algorithm: PNGA (Primary Nodes Group Algorithm). The objective of PNGA is promoting the robustness of the entire network. In the simulation experiment, this paper analyzes the network modeling and topological characteristics of a three-tier power grid in NV. We use the degree sorting algorithm as the control group which is widely used in power grid. Under different attack strategies, we investigate the performance of different algorithms and the state of fault generation. The results of the simulation we performed in this paper have shown that PNGA is superior to the rest of the algorithm in suppressing faults.

An attractive architecture called heterogeneous cloud radio access networks (H‐CRAN) becomes one of the important components of 5G networks, which can provide ubiquitous high‐bandwidth services with flexible network construction. However, massive access nodes increase the risk of cell outages, leading to negative impact on user‐perceived QoS (Quality of Service) and QoE (Quality of Experience). Thus, cell outage management (COM) became a key function proposed in SON (Self‐Organized Networks) use cases. Based on COM, cell outage detection (COD) will be resolved before cell outage compensation (COC). Currently few studies concentrate on COD for 5G H‐CRAN, and we propose self‐organized COD architecture and approach for it. We firstly summarize current COD solutions for LTE/LTE‐A HetNets and then introduce self‐organized architecture and approach suitable for H‐CRAN, which includes COD architecture and procedures, and corresponding key technologies for it. Based on the architecture, we take a use case with handover data analysis using modified LOF (Local Outlier Factor) detection approach to detect outage for different kinds of cells in H‐CRAN. Results show that the proposed approach can identify the outage cell effectively.

Hybrid cognitive radio (CR) relays serve cellular users in a two-hop fashion, which jointly utilize both licensed and unlicensed radio spectrums to significantly increase the system capacity. User equipments (UEs) need to be actively associated with the macro-cell BS or CR relays having a more lightly loaded spectrum if the quality of services (QoS) can be guaranteed. To this end, this paper investigates optimal user association for load balancing problem in cellular network with hybrid cognitive radio relays. Firstly, we propose a multi-objective user association optimization model to balance the loads among different tiers while reducing the total resource occupancy. Then, this multiobjective problem is converted into a single one by the linear weighing-sum method and a genetic algorithm is introduced to solve it. The numerical simulation results show that our proposed scheme can obtain more balanced resources occupation, better throughput performance, and lower blocking rate compared with the heuristic and max-power strategies.

The power SCADA system is designed to ensure the safe operation of the power system. The SCADA communication network as an information exchange carrier between remote terminal units and master stations, is the key part of the SCADA system, and it has a high requirement for security. However, due to the wide distribution of the network and the interconnected network structure, it is susceptible to risks. So there is an urgent need for accurate and real-time risk prediction. In this paper, we propose a risk prediction model based on entropy-gray model, where the gray model is used to predict the values of the network risk indexes, and the entropy method is to determine the weight of those risk indexes. Finally, the overall risk value of the network is decided with analytic hierarchy process. Simulation results show that the proposed entropy-gray method can achieve accurate and timely risk prediction.

The increasing traffic puts high demands on capacity for the next generation mobile networks. The millimeter-Wave (mmWave) communication system offers new opportunities to meet this requirement due to the tremendous amount of avail- able spectrum. However, the massive non-line-of-sight (NLOS) transmissions and the site constraints in urban environment are severely challenging the conventional way of deploying terrestrial low power nodes (LPNs). To address these problems, we introduce the mmWave aerial base station (mAeBS) in next generation mobile networks, which can be quickly and flexibly deployed to enhance the capacity in data traffic bursting areas. To maximize the enhancing effects, an ergodic capacity analytical model of mAeBS is proposed, considering both user distribution and environment conditions. Then an mAeBS 3D placement method based on the model is given. Simulation results show that the proposed method can achieve considerable capacity enhancement and supplement regional coverage as well.

Vulnerability assessment and optimization for wide area monitoring, protection and control system (WAMPAC) can enhance the robustness and sustainability of network. However, current assessment methods are incomplete and optimization methods ignore dynamic process. A comprehensive vulnerability assessment and optimization method is proposed. Firstly, for assessment, a comprehensive vulnerability indicator is designed to assess vulnerability of nodes and edges in the network integrating static and dynamic aspects. And then, to relieve unbalanced vulnerability distribution in the network, a routing optimization method is proposed by reconfiguring service routes on the edge with high vulnerability. Finally, the simulation is taken under a real system. Vulnerability assessment with the defined indicator is executed, and its correctness is proved as well. Then with the optimization method, the network vulnerability can be balanced, which takes on effective theoretical and practical significance.

Recently, densified small cell deployment with overlay coverage through Heterogeneous Networks (HetNets) has emerged as a viable solution for 5G mobile networks. Cell Outage Detection (COD) which is the essential functionality in Self-Organizing Network (SON) is designed to autonomously deal with unexpected faults. Typical methods for detecting cell outage are usually based on Manual Drive Tests (MDT). However, it is difficult to detect small cell outage by MDT measurements in HetNets, because the User Equipment (UE) served by these small cells can switch to the macro cell and keep the Reference Signal Received Power (RSRP) and Signal to Interference plus Noise Ratio (SINR) values normal. To resolve this issue, we propose a COD architecture based on the handover statistics. Our model concentrates on cell outage detection in a two-tier heterogeneous network. We process sequential handover statistics spatially and temporally in conjunction with data mining methods. Also, an improved LOF algorithm (M-LOF) is proposed to enhance the detection performance based on handover statistics. To evaluate the system performance, a set of tests has been carried out using some reasonable assumptions and network simulator we designed. The results of simulation show that our system is more effective to detect cell outage in comparison to the architecture using MDT measurements.

This paper proposes a risk assessment model of key service and optimization methods to reduce service risk in smart grid communication network. Firstly, we analyze the probability of failure of communication link and node which is induced by external factors, like natural disaster, human attack and system disturbances. Then using importance of services, links and nodes, we build the risk model of failure for key services. Further, we propose optimization methods based on Dijkstra algorithms to reduce the risk of key services. Finally, based on part of smart grid communication network topology structure from a Chinese province, the simulation results show that the risk of key services and whole network are reduced.

LTE-based Public Safety Network (PSN) is a wireless communication network which can provide efficient and reliable communication in disasters or emergencies for disaster relief and public protection. Therefore, ensuring that network congestion will not happen in PSN during an emergency is becoming increasingly important. LTE-based PSN is easy to be congested because part of spectrum resources is compressed to guarantee priority requirements of public safety users. In this paper, we develop a new method namely Selective Admission Control (SAC) mechanism to manage the radio bearers access to the commercial radio for Public Safety (PS) in LTE-based PSN. In the case of emergency, we select the traffic bearer with minimum estimated load increment accessing to the LTE-based PSN. The channel quality of new bearers should be taken into account, which means that in congestion, users who arrive earlier with poor channel quality will be rejected to reserve sufficient resources for users who arrive later with good channel quality. The simulation results show that the SAC mechanism can improve throughput by 36% and lower the rejection rate by 73% at most than reference method based on non-selective access control model, as a result effectively avoiding the network congestion and improving the utilization of spectrum resources for public safety communication.

This paper proposes a novel Gain-Aware Uplink-Downlink(GAUD) jointly resource allocation scheme to maximize the Device-to-Device(D2D) throughput while guaranteeing Quality of Service (QoS) of cellular users. We formulate the global optimization problem as a mixed integer nonlinear programming problem and decompose it into three sub-problems. Firstly, a method of jointly uplink and downlink reuse mode selection is proposed. Based on the throughout gain, each D2D pair is appropriately assigned by either downlink or uplink frequency resource to reuse. Then a heuristic scheduling is designed for fairness channel allocation in order to form D2D users as much as possible. At last, the Lagrangian dual algorithm is developed to solve the optimal power allocation. The simulation results show that our proposed jointly downlink-uplink resource reusing scheme can make the system throughput increased by about 35% and 50% higher than the scheme based on Only Downlink and Only Uplink resource reusing.

To realise load optimisation in cellular and public wireless local area network (WLAN) interworking network, a fairness preferred throughput maximisation (FPTM) optimisation model and a particular algorithm for it named joint UE‐AN association and resource allocation optimisation based on generalised benders decomposition are proposed in the study. The derived solution will give guidance on UE's access selection and resource allocation in cellular network to optimise the overall performance of the interworking network. Simulation results validate the performance on optimising access load in the interworking networks of FPTM model, which can practically enhance the effect of offloading from cellular network to WLAN and improve the total throughput.

Current research in base station(BS) energy consumption area is mostly devoted to the study of the static energy consumption or dynamic factor. But it lacks the research on the energy storage efficiency of the BS. In this paper, we propose a power modeling of BSs based on energy storage monitoring. Firstly, the system architecture and networking scheme of energy storage monitoring are presented, and the mathematical model of BS energy consumption is analyzed. Then, based on the current network's data, a large number of real data are collected through the energy consumption analysis system. We analyzed the effect of charge and discharge time on energy storage efficiency of power grid. Based on the monitoring data, we fit the model of energy consumption and get the model of overall power consumption. Finally, based on the use control template, the optimization scheme of reducing the power consumption of authority network is proposed, which has substantial economic and green value.

This paper proposes a novel indoor-outdoor caching relay system (CRS) and develops the corresponding caching mechanism, which can improve the utilization of wireless resources. It operates in two phases periodically. In Phase I, spectrum resources of the established links between MBS and user equipment (UE) are extracted to support data caching. In Phase II, caching relay system can directly serve indoor users and the fronthaul resources are released to serve other users. Simulations verify that compared with the conventional relay system (RS) the proposed CRS can improve the system throughput by at most 142% in the reusable data caching cases, only at the cost of temporarily suppressing the traffic rate to establish caching links.

As power communication network is more and more important in smart grid, to decrease the failure risk of power system caused by the interruption of communication service, this paper propose a novel routing optimization method based on risk predication for communication services. Firstly, we analyze the probability of failure of communication link and node which is induced by external factors, like winds and snows, equipment failures, and etc. Then based on importance of services, links and nodes, we calculate the risk of failure of communication link and node. Further, we propose three service risk indicators and corresponding improved Dijkstra algorithms to optimize service routing, thus to decrease the network failure probability. Finally, based on part of power grid topology structure from a Chinese province, the simulation results show that the service risk o and the risk of the whole network are also reduced.

With the rapid development of cellular network systems, the operators need more experience to deal with complicated network management system and wide range of Key Performance Indicators (KPIs). There are many indicators related to each other due to the definition or communication process. But several implicit associations still exist among these KPIs. This paper proposes an approach to figure out the implicit linear relationship among indicators clearly in which a new clustering technique is used for distinguishing different relationships. Data analysis using real network data shows that the approach can well divide data into clusters, and each cluster can effectively reflect the relationship between indicators.

Energy efficiency in future green cellular wireless networks poses a challenge to operators and researchers. An effective method of providing energy savings (ES) in base stations (BSs) is to switch off idle BSs or put them into sleep mode and subsequently migrate the traffic loads to active BSs in their neighborhood. The intrinsic issue related to such methods applied to heterogeneous networks (HetNets) is that the optimal selections of various types of BSs during both energy-saving and coverage-compensating (CC) processes are difficult to determine. In an attempt to resolve the insufficiency of existing strategies, we propose a novel traffic-aware self-organizing ES mechanism that enables efficient resource allocation and interference management in multi-level networks. To further develop optimal energy conservation procedures for BSs, a new constraint model is proposed. We employ coverage gaps and over-provisioning as the optimization objectives and analyze the effects of their weights. The performance in terms of energy savings is evaluated in an urban Long-Term Evolution (LTE) scenario with different types of BSs. The simulation results show that the proposed mechanism can maximize energy efficiency in heterogeneous cellular networks while guaranteeing the quality of service. This algorithm is autonomous in terms of decision making and execution.

This paper discusses how to assess video Quality of Experience (QoE) with image feature information which includes texture and saliency information. In order to compress and transmit the feature information, wavelet transform is conducted and the high-frequency component histograms are fitted using generalized Gaussian distribution. At end user side, the video distortion is measured by using Kullback-Leibler Divergence (KLD) and therefore MOS is evaluated using neural network fitting. The LIVE Video Quality Database is used for testing the performance of proposed method. result confirms that the proposed method is competitive and suitable for assessing the QoE of real-time video service.

Power models are crucial to assess the power consumption of base stations (BSs) without quantitive description. Currently available models seldom consider the dynamic factors such as indoor and outdoor temperature. As power model will affect the energy-saving gains of different green resolutions, in this paper we provide such power models for mobile communication BSs relying on practical data collected from several BSs with focus on dynamic factors, e.g., traffic load, indoor and outdoor temperature. The quantitative power models for communication equipment and air conditioning are defined and validated combined with the mathematical method of linear regression. With application of the models we develop an energy saving method which can save at least 10% of the power consumption per year through the simulation and analysis. Still, the method does not affect the normal operation of communication BSs, which takes on strong economy and green significance.

Reducing the energy consumption (EC) of base station (BS) is one of the major concerns in wireless cellular networks. Additionally, turning off some underutilized BSs during off-peak period and performing effective compensation without delay are the most efficient way to save energy. However, large-scale energy conservation yet remains to be investigated at macro level. In this paper, to solve the problem that long convergence time and poor convergence precision in the large-scale network, we propose a BS topology-aware based energy-saving (ES) model, whose core is cell adjacency graph (CAG) with vertexes and links representing eNodeBs (eNBs) and their neighboring relationship. In addition, we introduce new metrics, predicted energy efficiency (PEE) and quality of compensation (QoC), as the weights of nodes and links respectively. Consequently, the model transforms the ES problem into average weights maximization in CAG. In view of the model presented, centralized and hybrid algorithms are put forward to solve the problem. Compared with classic distributed algorithm, simulation results claim that our hybrid approach achieves the maximization of ES with guaranteed QoC while our centralized approach maximize the PEE.

It's a challenge to effectively assess Quality of Experience (QoE) for TCP video streaming with network performance parameters, to resolve this problem, an objective hierarchical Evaluation for Transmission Control Protocol (TCP) video streaming is proposed under video playback scenarios. QoE assessment for TCP video streaming is resolved into two sub-steps. In the first place, in consideration of video playback performance parameters affecting QoE, the authors demonstrate three novel application-layer metrics. Further, impact of network status on video playback performance is investigated and the authors propose high level network-layer parameters. Then the correlation between the network-layer parameters and application-layer metrics is characterized through analysis and inference. In the secondly place, subjective tests are conducted to evaluate QoE from application-layer metrics. Ultimately, the authors validate analysis and model by simulations and experiments in real networks. The experimental study shows that the proposed method performs well in assessing QoE of TCP video streaming.

Load balancing is one of the key target of LTE Self-Optimization Network (SON). In this paper, we propose a load balancing method for LTE downlink network, namely Load Vector Minimization based Load Balancing (LVMLB) method. Load Vector (LV) is a vector whose elements are the load values of cells and sorted in descending order. The order of LVs is defined by the lexicographical order. The smaller the LV is, the higher the balance degree of cells load will be. As the LV has a lower bound with total load fixed, the balance degree of cells load would reach a local optimal. On this basis, we design the LVMLB algorithm, trying to get the optimal solutions to load balancing problems, the proof of being optimal will also be given in this paper. Simulation scenarios are set in a square part of Macro-Pico mixed HetNets. Simulation results show that LVMLB outperforms the Cell Region Expansion (or Bias) scheme, increasing the capacities of Macro and Pico tiers at the same time, and improving balance degree of cells load, only sacrificing a little QoS performance.

The current power consumption of intelligent terminals are over burden for their battery capacities, which directly restrict the hours used. In order to realize the energy saving of terminals in the LTE-A system, this paper puts forward the concept of virtual cells and a related uplink energy saving mechanism. Virtual Cell's resources and the outage probability of terminals are proposed by this mechanism as constraint conditions. The first step is sectioning off energy saving area in virtual cell. Secondly, we use the uplink CoMP (Coordinated Multiple Points Transmission/Reception) technology to provide diversity gain for the terminals of energy saving area. The third step depends on uplink power control which could adjust the mobile terminals' transmission power for energy saving. The simulation results show that the energy consumption of total terminals will decrease nearly 50% in the virtual cell while its capacity is lower than the 50% of maximum.

A novel dynamic multi-stage ESM (Energy-Saving Management) mechanism based on BS (Base Station) cooperation is proposed. The mechanism firstly introduces a local OP (Opposite Pair) cooperation method taking account of geographic topology and then divides time period into four domains. In time domains, regional dynamic multi-stage algorithms and efficient performance evaluation model for the mechanism is analyzed as well. The mechanism is simulated under a practical LTE BS deployment. Results show that 25.1% of regional energy can be saved at most. Still better coverage, interference, and throughput performance can be obtained comparing to other algorithms.

For the sake of the increase of energy consumption, the emission of carbide can't be negligible. In this paper, we propose a mathematical model which adjusts the radius of base station, to figure out the problem of energy saving, while the quality of services such as the traffic volume of each base station and the service area should be satisfied. For the model, we use ant colony algorithm as a practical method to solve this problem. By using this mechanism, we can achieve the purpose of energy saving, when the traffic in a service area is low.

To solve the abnormal coverage problems caused by unreasonable network parameter settings more effectively, an automated coverage optimization scheme based on downtilt-adjustment of base stations in wireless access networks is proposed. After detecting and analyzing the abnormal coverage situation, simulated annealing algorithm is adopted by the scheme to figure out an optimal downtilt-adjustment solution for each base station. And then each base station can effectively adjust its electronic downtilt according to the solution to optimize the coverage. The whole process is completed without human intervention. Simulation results show that the proposed automated coverage optimization scheme can improve the wireless network coverage quality. Moreover, weak coverage and excessive coverage problems can be solved effectively.

Currently, the way that computers communicate has been changed, such as the increased management cost and the decreased performance. In order to resolve the problem, IT companies often outsource their part function to service providers. Many services require a similar infrastructure, therefore, we design a generic resource pool in which many services can share these resources. For this purpose, the first step is to propose a new dynamic and distributed algorithm for allocating resources to a set of services. Additionally, the proposed algorithm takes into account server resources (CPU and memory) and network related resources (delay and bandwidth), and the distributed algorithm takes CPU and bandwidth as the variables for the objective function. And it aims at maximizing the satisfied demand. Based on the distributed algorithm, the percentage of the satisfied demand is maximized. New algorithm is then used to do service placement and server selection.