Reliable Nodes Research Articles

Hub Node Identification in Urban Rail Transit Network Evolution Using a Ridership-Weighted Network

With the development of the urban rail transit network (URTN), the network structure and performance have changed, and the node importance has also been redistributed. However, little research has been done on how hub nodes change as the network develops over a lengthy period. Moreover, most hub node identification methods only focus on the analysis of topological networks or single-dimension measurements, resulting in inaccurate identification results. To overcome the above limitations, a novel method of hub node identification is proposed. Based on the ridership-weighted network model, the node centrality and reliability are aggregated to quantify the weighted comprehensive importance of the nodes. Furthermore, network invulnerability measurement is used to demonstrate the effectiveness of the proposed method. This method is applied to the Xi’an Urban Rail Transit Network (XURTN) from 2011 to 2021. With the XURTN’s development, its connectivity, balance, and fault tolerance have improved. After the basic network skeleton was formed, the number and proportion of hub nodes increased steadily. By comparing the spatial characteristics of the identified hub nodes over two successive periods, it can be found that the evolution direction of the hub nodes is correlated with the type of new lines and coincides also with the development direction of the urban area. In addition, the node orders of the proposed method have a greater impact on the network vulnerability, in which the network-weighted efficiency [Formula: see text] decreases faster and more dramatically, that is, 1.17%–45.75% more than that of other methods. Overall, this study provides a basis for the URTN and station planning and management.

Contrastive learning of graphs under label noise

In the domain of graph-structured data learning, semi-supervised node classification serves as a critical task, relying mainly on the information from unlabeled nodes and a minor fraction of labeled nodes for training. However, real-world graph-structured data often suffer from label noise, which significantly undermines the performance of Graph Neural Networks (GNNs). This problem becomes increasingly severe in situations where labels are scarce. To tackle this issue of sparse and noisy labels, we propose a novel approach Contrastive Robust Graph Neural Network (CR-GNN), Firstly, considering label sparsity and noise, we employ unsupervised contrastive loss and further incorporate homophily in the graph structure, thus introducing neighbor contrastive loss. Moreover, data augmentation is typically used to construct positive and negative samples in contrastive learning, which may result in inconsistent prediction outcomes. Based on this, we propose a dynamic cross-entropy loss, which selects the nodes with consistent predictions as reliable nodes for cross-entropy loss and benefits to mitigate the overfitting to labeling noise. Finally, we propose cross-space consistency to narrow the semantic gap between the contrast and classification spaces. Extensive experiments on multiple publicly available datasets demonstrate that CR-GNN notably outperforms existing methods in resisting label noise.

Deep neural network with fuzzy algorithm to improve power and traffic-aware reliable reactive routing

<span>In wireless networks, link breaks, and restricted resources create fundamental challenges for maintaining network applications. Several wireless network routing techniques concentrate on power efficiency to expand the network lifetime, but the traffic and reliability parameters are not the primary concern. Though, these techniques are not capable of dealing with the wireless network. Hence, this paper proposes deep neural network (DNN) with a fuzzy algorithm to improve power and traffic-aware reliable reactive routing (PTAR) in wireless networks. The wireless network is formed by clustering by the node power and selects the cluster head (CH) based on a fuzzy algorithm. The wireless node power level, node buffer space, and node reliability to consider the input parameters of the fuzzy system. Then thefuzzy algorithm gives the output for CH round length. This selected CH improves the node reliability, power efficiency with minimized network congestion. Then we use a DNN algorithm to choose an optimal relay by applying an adaptive load balance factor in the network. DNN is a machine learning algorithm, and it provides high accuracy. From the simulation results, the PTAR approach improves the network performance, such as packet received ratio, delay, residual energy, and routing overhead.</span>

Методика розрахунку основних характеристик підшипників ковзання систем і агрегатів автомобілів та аналізу їх трибологічної ефективності

A methodology for calculating the main characteristics of sliding bearings of car systems and units is proposed. The specified methodical approach is important when evaluating indicators of tribological efficiency of couplings of parts, nodes, systems and aggregates and operational reliability of cars as a whole. The bearing activation model makes it possible to formulate measures to increase the tribological efficiency of the couplings of parts and the level of operational reliability of nodes, systems and aggregates, and cars as a whole. In addition, the parameters of the actuation model also respond to friction modes. It is shown that it is possible to increase the tribological efficiency of the sliding bearing by means of design, technological and operational methods. The main formulas that make it possible to develop algorithms for calculating a set of bearing characteristics are presented: contact pressure, stress, strength, sliding resistance, wear, resource. For this purpose, the following values are used: compliance coefficients and relative load, contact angle, contact pressures, contact displacements, and the stress state in the contact zone. The geometric and power parameters of the bearing are involved in the developed methodology and algorithm. Modes of friction and the nature and magnitude of the load on the bearing are also taken into account. An algorithm for calculating auxiliary and main characteristics of a sliding bearing has been developed. The conditions of continuity and equilibrium in contact are clarified. The differential-integral equation of the operation of the bearing is obtained and the dependence of the current contact angle on the friction path and other parameters of the sliding bearing operation model is obtained. It is shown that the tribological efficiency of sliding bearings is established by the method of comparative analysis of tribotechnical characteristics determined by the results of tests on friction machines and a stand.

An efficient secure and energy resilient trust-based system for detection and mitigation of sybil attack detection (SAN).

In the modern digital market flooded by nearly endless cyber-security hazards, sophisticated IDS (intrusion detection systems) can become invaluable in defending against intricate security threats. Sybil-Free Metric-based routing protocol for low power and lossy network (RPL) Trustworthiness Scheme (SF-MRTS) captures the nature of the biggest threat to the routing protocol for low-power and lossy networks under the RPL module, known as the Sybil attack. Sybil attacks build a significant security challenge for RPL networks where an attacker can distort at least two hop paths and disrupt network processes. Using such a new way of calculating node reliability, we introduce a cutting-edge approach, evaluating parameters beyond routing metrics like energy conservation and actuality. SF-MRTS works precisely towards achieving a trusted network by introducing such trust metrics on secure paths. Therefore, this may be considered more likely to withstand the attacks because of these security improvements. The simulation function of SF-MRTS clearly shows its concordance with the security risk management features, which are also necessary for the network's performance and stability maintenance. These mechanisms are based on the principles of game theory, and they allocate attractions to the nodes that cooperate while imposing penalties on the nodes that do not. This will be the way to avoid damage to the network, and it will lead to collaboration between the nodes. SF-MRTS is a security technology for emerging industrial Internet of Things (IoT) network attacks. It effectively guaranteed reliability and improved the networks' resilience in different scenarios.

Detection of E-Commerce Fraud Review via Self-Paced Graph Contrast Learning

Abstract Recently, graph neural networks (GNNs) have been widely used for e-commerce review fraud detection by aggregating the neighborhood information of nodes in various relationships to highlight the suspiciousness of nodes. However, existing GNN-based detection methods are susceptible to sample class imbalance and fraud camouflage problems, resulting in poor quality of constructed graph structures and inability to learn reliable node embeddings. To address the above problems, we propose a novel e-commerce review fraud detection method based on self-paced graph contrast learning (SPCL-GNN). Firstly, the method constructs a subgraph by initially selecting nodes through a labeled balanced extractor. Secondly, the subgraph connections are filtered and complemented by combining self-paced graph contrast learning and an adaptive neighbor sampler to obtain an optimized graph structure. Again, an attention mechanism is introduced in intra- and inter-relationship aggregation to focus on the importance of aggregation under different relationships. Finally, the quality of the node embedding representation is further improved by maximizing the mutual information between the local and global representations. Experimental results on the Amazon and YelpChi datasets show that SPCL-GNN significantly outperforms the baseline.

Label Efficient Regularization and Propagation for Graph Node Classification.

An enhanced label propagation (LP) method called GraphHop was proposed recently. It outperforms graph convolutional networks (GCNs) in the semi-supervised node classification task on various networks. Although the performance of GraphHop was explained intuitively with joint node attribute and label signal smoothening, its rigorous mathematical treatment is lacking. In this paper, we propose a label efficient regularization and propagation (LERP) framework for graph node classification, and present an alternate optimization procedure for its solution. Furthermore, we show that GraphHop only offers an approximate solution to this framework and has two drawbacks. First, it includes all nodes in the classifier training without taking the reliability of pseudo-labeled nodes into account in the label update step. Second, it provides a rough approximation to the optimum of a subproblem in the label aggregation step. Based on the LERP framework, we propose a new method, named the LERP method, to solve these two shortcomings. LERP determines reliable pseudo-labels adaptively during the alternate optimization and provides a better approximation to the optimum with computational efficiency. Theoretical convergence of LERP is guaranteed. Extensive experiments are conducted to demonstrate the effectiveness and efficiency of LERP. That is, LERP outperforms all benchmarking methods, including GraphHop, consistently on five common test datasets, two large-scale networks, and an object recognition task at extremely low label rates (i.e., 1, 2, 4, 8, 16, and 20 labeled samples per class).

Swarm Intelligence-Optimized Energy Management for Prolonging the Lifetime of Wireless Sensor Networks

Recent technological and industrial progress has enabled the development of small, high-performing, energy-saving, affordable sensor nodes that possess the potential to adapt, be self-aware, and self-organize. These nodes are designed for versatile communications applications. Sensor networks for sustainable development focus on the ways in which sensor network technology can enhance social development and improve living standards without causing harm to the environment or depleting natural resources. Wireless sensor networks (WSNs) offer undeniable benefits in various fields, including the military, healthcare, traffic monitoring, and remote image sensing. Given the constraints of sensor networks, varying degrees of security are necessary for these critical applications, posing difficulties in the implementation of conventional algorithms. The issue of security has emerged as a primary concern in the context of IoT and smart city applications. Sensor networks are often regarded as the fundamental building blocks of IoTs and smart cities. The WSN encompasses a routing algorithm, network strength, packet loss, energy loss, and various other intricate considerations. The WSN also addresses intricate matters such as energy usage, a proficient approach for picking cluster heads, and various other concerns. The recent growth of Wireless Sensor Networks (WSNs) has made it increasingly difficult to ensure the trustworthiness and reliability of data due to the distinct features and limitations of nodes. Hostile nodes can easily damage the integrity of the network by inserting fake and malicious data, as well as launching internal attacks. Trust-based security is employed to detect and identify rogue nodes, providing a robust and adaptable protection mechanism. Trust evaluation models are crucial security-enhancement mechanisms that enhance the reliability and collaboration of sensor nodes in wireless sensor networks. This study recommends the use of DFA UTrust, a unique trust technique, to effectively satisfy the security requirements of WSNs.

Dingo optimization influenced arithmetic optimization – Clustering and localization algorithm for underwater acoustic sensor networks

Underwater acoustic sensor networks (UASNs) are quickly being used in environmental monitoring, marine life observation, and underwater resource research. This paper introduces the Dingo Optimisation Influenced-Arithmetic Clustering and Localisation Algorithm. The proposed method was developed due to the above considerations. DOIAO-CLA combines Dingo Optimisation (DO) and Arithmetic Optimisation (AO), two powerful optimisation methods. DOIAO-CLA uses DO's robust exploration and exploitation features to group sensor nodes with similar propagation characteristics. It also optimises cluster formation and reduces distance discrepancies using AO. DOIAO-CLA also enables reliable localization by merging distance- and angle-based methods. DO optimises node localization by using distance and angle of arrival data. This integration enhances UASN node localization accuracy and reliability. The DOIAO-CLA was tested against other clustering and localization algorithms like Archimedes Optimisation Algorithm (AOA), Squirrel Search Algorithm (SSA), Hybrid Cat Cheetah Optimisation Algorithm (HCCOA), Ant Colony Optimisation Algorithm (ACOA), Grey Wolf Optimisation Algorithm (GWOA), and Particle Swarm Optimisation Algorithm. According to the results, DOIAO-CLA taps alternatives for network robustness, data aggregation, and localization precision. Finally, DOIAO-CLA solves the UASN clustering and localization problem in a novel and effective way.

Enabling large-scale low-power LoRa data transmission via multiple mobile LoRa gateways

Low-Power Wide Area Networks (LPWANs) enable large-scale Internet of Things (IoT) applications. LoRa is a promising LPWAN technology, but LoRa nodes are battery-powered, so their lifetime depends on energy efficiency. In practice, LoRa nodes have short battery life because long transmission distances to gateways incur high energy costs. With expensive battery replacement, particularly in large networks, it is crucial to reduce LoRa nodes’ energy consumption. Existing low-power transmission techniques focus on static or single mobile gateways, consuming substantial remote node energy unfit for large-scale networks. This paper proposes integrating LoRa with mobility to minimize node energy use by shortening transmission distances. We design the first large-scale mobile LoRa system, MLoRaDrone, using multiple unmanned aerial vehicles (UAVs) equipped with LoRa gateways flying near nodes. First, we present a low-power mechanism for reliable node sensing and UAV communication. Next, we develop a LoRa transmission parameter assignment strategy encompassing a distributed ADMM-based optimal transmission policy and a 32-approximation channel allocation. Moreover, we propose a UAV trajectory scheduling scheme with a 4-approximation minimizing node energy and UAV flight paths. Evaluations on varied scales verify MLoRaDrone effectiveness for different node counts and UAV paths. Compared to baselines, MLoRaDrone reduces 3000-node energy consumption by up to 35.56× within 25 km2.

Transmission reliability evaluation of wireless sensor networks considering channel capacity randomness and energy consumption failure

The reliable transmission of data packets is the basis of ensuring the Internet of Things applications. Therefore, how to ensure the transmission reliability of wireless sensor networks has already become a hotspot. However, the existing researches ignore the effects of multiple fault modes and random factors on transmission reliability. Firstly, in the node reliability modeling, few studies have considered the effects of random failure and energy consumption failure comprehensively. Secondly, in the link reliability modeling, existing studies ignore the effect of the channel capacity randomness, which can lead to the erroneous estimation of transmission reliability. To overcome the above shortcomings, we propose a more realistic transmission reliability model. Firstly, considering the comprehensive effects of two failure modes, we establish a new node reliability model to better characterize the node performance. Secondly, considering the channel capacity randomness, we propose an improved SNR-Capacity connection model by deriving the one-hop connectivity probability, which can better characterize the link reliability. Then, we propose a transmission reliability evaluation model, and give the corresponding exact and approximate algorithms for its solution. Finally, taking the military Internet as an example, the effectiveness of the proposed method is verified.

Self-Healing Secure Blockchain Framework in Microgrids

Blockchain has recently been depicted as a secure protocol for information exchange in cyber-physical microgrids. However, it is still found vulnerable to consensus manipulation attacks. These stealth attacks are often difficult to detect as they use kernel-level access to mask their actions. In this paper, we firstly build a trusted and secured peer-to-peer network mechanism for physical DC microgrids’ validation of transactions over Distributed Ledger. Secondly, we leverage from a physics-informed approach for detecting malware-infected nodes and then recovering from stealth attacks using a self-healing recovery scheme augmented into the microgrid Blockchain network. This scheme allows compromised nodes to adapt to a reconstructed trustworthy signal in a multi-hop manner using corresponding measurements from the reliable nodes in the network. Additionally, recognizing the possible threat of denial-of-service attacks and random time delays (where information sharing via communication channels is blocked), we also integrate a model-free predictive controller with the proposed system that can locally reconstruct an expected version of the attacked/delayed signals. This supplements the capabilities of Blockchain, enabling it to detect and mitigate consensus manipulation attempts, and network latencies.

A comprehensive review on trust management approaches in fog computing

Decentralised fog computing can provide real-time interaction, minimize latency, heterogeneity, and provide networking services between edge devices and cloud data centers. One of the biggest challenges in the fog layer network is finding a trustworthy fog node. Trust management encompasses the process of being trustworthy and the act of assessing the reliability of other nodes. It is essential to carry out a comprehensive review using a systematic approach in this field to advance our understanding, address emerging challenges, and foster secure and efficient trust management practices. This research paper considers a comprehensive analysis of high-quality fog computing trust management literature from 2018 to 2022. A variety of distinct approaches have been chosen by fog computing-based trust management and these techniques are classified into three categories: algorithms, challenges, and limitations. Further, it reviews the various trust attacks in fog environments, details the solutions proposed in the current literature, and concludes with a discussion of the open challenges and potential future research directions in fog computing.

Learning node representations against perturbations

Recent graph neural networks (GNN) has achieved remarkable performance in node representation learning. One key factor of GNN’s success is the smoothness property on node representations. Despite this, most GNN models are fragile to the perturbations on graph inputs and could learn unreliable node representations. In this paper, we study how to learn node representations against perturbations in GNN. Specifically, we consider that a node representation should remain stable under slight perturbations on the input, and node representations from different structures should be identifiable, which two are termed as the stability and identifiability on node representations, respectively. To this end, we propose a novel model called Stability-Identifiability GNN Against Perturbations (SIGNNAP) that learns reliable node representations in an unsupervised manner. SIGNNAP formalizes the stability and identifiability by a contrastive objective and preserves the smoothness with existing GNN backbones. The proposed method is a generic framework that can be equipped with many other backbone models (e.g. GCN, GraphSage and GAT). Extensive experiments on six benchmarks under both transductive and inductive learning setups of node classification demonstrate the effectiveness of our method. Codes and data are available online: https://github.com/xuChenSJTU/SIGNNAP-master-online

Reengineering and Its Reliability: An Analysis of Water Projects and Watershed Management under a Digital Twin Scheme in China

Water project and watershed management is currently being reengineered under digital twin schemes in China through pilot projects. An evaluation on pilot reengineering is important for its further implementation and improvement. This paper investigates national legislation and pilot projects’ implementations of a digital twin watershed and digital twin water project from the perspectives of design, policy, technology, investment, personnel, cyberspace security, co-construction, and sharing through interviews and expert review, and it uses a Bayesian network to study their reliability. First, the design of the digital twin watershed and digital twin water project is reasonable with regard to system architecture and business management. Second, although there are some national legislations on cyberspace security and geospatial data, they are incomplete for policy making and are probably infeasible for some technology. Third, there are insufficient mechanisms to sustainably support investment, personnel, and cyberspace security. Forth, co-construction and sharing are required for both inside and outside water departments. Fifth, the Bayesian network is useful for investigating the reliability of weak nodes, and it is helpful for the design and further implementation of the digital twin watershed and digital twin water project, as will be demonstrated with an anonymous example. This study could provide useful insights into the further reengineering of water projects and watershed management under a digital twin scheme in the world.

A novel fuzzy trust-based secure routing scheme in flying ad hoc networks

Today, many studies assess vulnerabilities, threats, and attacks in flying ad hoc networks (FANETs) to provide solutions for countermeasures. Protecting FANETs against attackers and coordinating connections are challenging. The purpose of this study is to increase and maintain communication security. In this paper, a fuzzy trust-based secure routing scheme (FTSR) is presented in FANETs. FTSR utilizes two trust assessment mechanisms, namely local trust and path trust. Local trust strategy is a distributed process for finding reliable neighboring nodes and isolating hostile nodes on the network. In this regard, only reliable nodes are allowed to contribute to the path discovery procedure. This lowers the risk of forming fake paths in FANETs. Path trust strategy is responsible for identifying hostile nodes that are not identified in the local trust process. This strategy shows a general view of the trust status of the desired path. To design this mechanism, the source node runs a fuzzy system to select the safest path between source and the destination. Finally, network simulator 2 (NS2) implements FTSR, and the results such as malicious detection rate, packet delivery ratio, packet loss, accuracy, and delay are obtained from the simulation process. These results indicate that FTSR presents better performance compared to TOPCM, MNRiRIP, and MNDA. However, FTSR takes more time to find paths compared to TOPCM.

A solution for resource allocation through complex systems in fog computing for the internet of things

To manage the IoT resources to process the massive amount of collected data in complex systems, Allocation of tasks to nodes to increase user satisfaction has been noticed by researchers. In allocating tasks, the necessity of a powerful auxiliary platform is felt. Cloud and fog networks are two suitable options for performing and completing tasks on the Internet of Things. In the cloud network, due to offloading tasks to remote computing centers, problems such as increasing complete time, increasing traffic in the network, and high energy consumption for the Internet of Things network occur. The fog network solves the problem of completing tasks due to reducing the geographical distance between Internet of Things devices and processor nodes. The devices in the fog network cooperate with the Internet of Things network’s diverse and variable nodes with time computing, storage, and processing resources. In this heterogeneous environment, the non-cooperation and dishonesty of malicious fog nodes through misrepresentation of shared resources and non-completion of tasks can disrupt the performance of the Internet of Things network. As a result, two fundamental challenges in the heterogeneous fog network include providing resources to allocate tasks and recognizing valid fog nodes. This paper focuses on creating a resource provider network to allocate tasks to authentic fog nodes with appropriate resources. When an Internet of Things node faces a shortage of resources and link prediction, it can form a resource supply network with neighboring nodes to solve the two stated challenges. For this purpose, in the first stage, efforts are made to select the most appropriate options among the neighboring nodes based on specific criteria. In the second step, reliable nodes are identified based on a trust mechanism, including resource trust and entity trust. In the last step, tasks are divided among reliable devices with sufficient resources. The evaluation of the proposed method in the OMNET++ simulator shows that the trust mechanism can lead to the identification of reliable sources and, as a result, reduce the time to complete tasks.

A Trust-Based Reliable Confident Information Coverage Model of Wireless Sensor Networks for Intelligent Transportation

Wireless sensor nodes are vulnerable to interference or intrusion, especially in the tunnel scenario of intelligent transportation, which leads to potential risks to network security and reliability. Reliable coverage can ensure data perception and transmission, which helps to strengthen the reliability and quality of service (QoS) of wireless sensor networks (WSNs). Existing coverage models, such as the disc model, do not consider malicious intrusion, which still has a great gap with the realistic application scenarios. Therefore, it is urgent to study a coverage model considering network intrusion and node reliability. In this paper, a novel trust-based reliable confident information coverage (T-RCIC) model is proposed based on the confident information coverage (CIC) model and trust management. Trust management is adopted to determine whether a node is trusted, critical trusted, or malicious. If a node is identified as malicious, all its functions will be blocked. For critical trusted nodes, the trust value is utilized to determine the trust feature distance to reduce its proportion of coverage estimation. Furthermore, the reliable coverage degree (RCD) is defined to evaluate coverage reliability from node trust value and reliable coverage percentage. Simulation results indicate that the proposed model inherits the advantages of the CIC model in the spatial correlation and can effectively identify malicious nodes. The proposed T-RCIC model can significantly enhance coverage reliability.

Weight-Based PA-GPSR Protocol Improvement Method in VANET.

Vehicle Ad-hoc network (VANET) can provide technical support and solutions for the construction of intelligent and efficient transportation systems, and the routing protocol directly affects the efficiency of VANET. The rapid movement of nodes and uneven density distribution affect the routing stability and data transmission efficiency in VANET. To improve the local optimality and routing loops of the path-aware greedy perimeter stateless routing protocol (PA-GPSR) in urban sparse networks, a weight-based path-aware greedy perimeter stateless routing protocol (W-PAGPSR) is proposed. The protocol is divided into two stages. Firstly, in the routing establishment stage, the node distance, reliable node density, cumulative communication duration, and node movement direction are integrated to indicate the communication reliability of the node, and the next hop node is selected using the weight greedy forwarding strategy to achieve reliable transmission of data packets. Secondly, in the routing maintenance stage, based on the data packet delivery angle and reliable node density, the next hop node is selected for forwarding using the weight perimeter forwarding strategy to achieve routing repair. The simulation results show that compared to the greedy peripheral stateless routing protocol (GPSR), for the maximum distance-minimum angle greedy peripheral stateless routing (MM-GPSR) and PA-GPSR protocols, the packet loss rate of the protocol is reduced by an average of 24.47%, 25.02%, and 14.12%, respectively; the average end-to-end delay is reduced by an average of 48.34%, 79.96%, and 21.45%, respectively; and the network throughput is increased by an average of 47.68%, 58.39%, and 20.33%, respectively. This protocol improves network throughput while reducing the average end-to-end delay and packet loss rate.

GraphSR: A Data Augmentation Algorithm for Imbalanced Node Classification

Graph neural networks (GNNs) have achieved great success in node classification tasks. However, existing GNNs naturally bias towards the majority classes with more labelled data and ignore those minority classes with relatively few labelled ones. The traditional techniques often resort over-sampling methods, but they may cause overfitting problem. More recently, some works propose to synthesize additional nodes for minority classes from the labelled nodes, however, there is no any guarantee if those generated nodes really stand for the the corresponding minority classes. In fact, improperly synthesized nodes may result in insufficient generalization of the algorithm. To resolve the problem, in this paper we seek to automatically augment the minority classes from the massive unlabelled nodes of the graph. Specifically, we propose \textit{GraphSR}, a novel self-training strategy to augment the minority classes with significant diversity of unlabelled nodes, which is based on a Similarity-based selection module and a Reinforcement Learning(RL) selection module. The first module finds a subset of unlabelled nodes which are most similar to those labelled minority nodes, and the second one further determines the representative and reliable nodes from the subset via RL technique. Furthermore, the RL-based module can adaptively determine the sampling scale according to current training data. This strategy is general and can be easily combined with different GNNs models. Our experiments demonstrate the proposed approach outperforms the state-of-the-art baselines on various class-imbalanced datasets.