关键词： Prototypes   Computational modeling   Data models   Malware   Adaptation models   Servers   Federated learning   Convergence   Training   Data privacy   Mobile edge computing   personalized federated learning   non-IID data   label-free prototypes   contrastive learning   Android malware recognition  

摘要： Mobile Edge Computing (MEC) facilitates computing and storage at edge nodes near user devices, reducing latency and optimizing bandwidth. Federated Learning (FL) complements MEC by enabling privacy-preserving collaborative model training across edge nodes without sharing raw data. However, in MEC environments, FL faces challenges such as communication inefficiency and data heterogeneity (Non-IID), which degrade model performance and hinder convergence. To address these issues, we propose FedLFP, a communication-efficient personalized federated learning approach using label-free prototypes for Non-IID data in MEC. FedLFP employs three key strategies: (1) a Label-Free Prototype strategy to reduce communication costs and mitigate privacy risks, (2) a centroid prototype and combined clustering weight strategy to improve global prototype quality by considering data quantity and confidence levels, and (3) a multifaceted weighted contrastive learning strategy to enhance local representation learning and global alignment. We evaluated FedLFP on Android malware recognition using the KronoDroid dataset and standard image classification tasks, with eight configurations representing practical Non-IID settings. Experimental results show that FedLFP consistently outperforms thirteen state-of-the-art FL methods in accuracy, communication and computational efficiency. Additionally, we provide theoretical guarantees for the convergence of FedLFP under Non-IID conditions.

关键词： Mobility management   Handover decision   Mobile edge computing   Machine learning   Heterogeneous networks   6G network  

摘要： The forthcoming Six-Generation (6G) cellular network promises to provide novel developments in network architecture, offering greater data rates and ultra-low latency, ensuring seamless and reliable connectivity with a high quality of service for a massive number of connected devices. Even though efficient handover decisionmaking is one of the critical challenges in 6G networks, especially with high mobility scenarios and complex characterization of future networks. The case becomes more critical with the implementation of Mobile Edge Computing (MEC), which will lead to making the handover decision process more challenging due to its characterization and high requirements. This paper presents a systematic review of the handover decision (HOD) with MEC in 6G networks, providing a deep understanding of the most standing challenges and solutions addressing mobility management issues in 6G mobile networks. Moreover, machine learning (ML) and deep learning (DL) technologies are the key promising solutions for intelligent HOD-making in 6G networks with MEC. Therefore, this research work also aims to give a main focus on studying and highlighting the advanced ML methods that can be used to enhance HOD-making in 6G cellular networks with MEC. Furthermore, a comprehensive review of HOD-based ML solutions is provided to enhance the Quality of Service (QoS) of user experience in Heterogeneous Networks (HetNet), instilling confidence in the paper's findings. Besides, proposed solutions for HODs using ML models with next-generation network requirements and possible technologies are presented. We also describe research challenges and future directions for achieving this study.

关键词： Resource management   Pricing   Servers   Heuristic algorithms   Dynamic scheduling   Optimization   Games   Energy consumption   Vehicle dynamics   Simulation   DISC-PSO algorithm   dynamic pricing   elastic edge offloading   mobile edge computing   near-optimal resource allocation  

摘要： In mobile edge computing (MEC), task offloading can significantly reduce task execution latency and energy consumption of end user (EU). However, edge server (ES) resources are limited, necessitating efficient allocation to ensure the sustainable and healthy development for MEC system. In this paper, we propose a dynamic pricing mechanism based near-optimal resource allocation for elastic edge offloading. First, we construct a resource pricing model and accordingly develop the utility functions for both EU and ES, the optimal pricing model parameters are derived by optimizing the utility functions. In the meantime, our theoretical analysis reveals that the EU's utility function reaches a local maximum within the search range, but exhibits barely growth with increased resource allocation beyond this point. To this end, we further propose the Dynamic Inertia and Speed-Constrained particle swarm optimization (DISC-PSO) algorithm, which efficiently identifies the near-optimal resource allocation. Comprehensive simulation results validate the effectiveness of DISC-PSO algorithm, demonstrating that it significantly outperforms existing schemes by reducing the average number of iterations to reach a near-optimal solution by 86.88%, increasing the EU utility function value by 0.13%, and decreasing the variance of results by 96.78%.

关键词： Federated learning   Cooperative game   Resource allocation   Mobile edge computing   Internet of things  

摘要： Hierarchical federated learning has emerged as a promising approach to enhance the scalability and efficiency of federated learning in large-scale Internet of Things (IoT) environments. However, existing frameworks often struggle with energy efficiency, adaptability to dynamic conditions, and the complex interplay between device selection and resource allocation. In this paper, we propose an energy-efficient hierarchical federated learning framework that jointly optimizes IoT device selection, local resource allocation, and base station resource allocation across a three-layer architecture. We decompose the mixed-integer non-linear programming problem into two subproblems and propose an iterative algorithm that alternates between solving these subproblems until convergence. For the IoT device selection and local resource allocation subproblem, we introduce a multi-agent actor-critic approach enabling autonomous decision-making at the device level. The base station resource allocation subproblem is addressed using a cooperative game-theoretic approach, facilitating distributed coordination among base stations. Our simulation results demonstrate significant improvements in energy efficiency, learning performance, and adaptability compared to baseline methods.

关键词： Autonomous aerial vehicles   NOMA   Trajectory   Servers   Resource management   Optimization   Decoding   Reflection   Symmetric matrices   Communication system security   Intelligent reflective surface   mobile edge computing   non-orthogonal multiple access   autonomousaerial vehicle  

摘要： To cope with the challenges of traditional mobile edge computing (MEC), such as inflexibility and limited coverage, autonomousaerial vehicle (AAV)-MEC has been proposed as an important paradigm. However, the AAV-MEC systems are still facing the severe channel fading and obstacle blocking, which are fortunately expected to be solved by the emerging intelligent reflective surface (IRS) technique. In this paper, we propose an IRS-assisted AAV-MEC framework, where a AAV employs non-orthogonal multiple access (NOMA) scheme to offload partial computation tasks to multiple ground MEC servers connecting to access points (APs) with the assistance of IRS. The problem of maximizing computational capacity of the network is proposed by jointly optimizing the phase shift of IRS, transmit power and computation resource allocation of AAV, NOMA decoding order, and AAV trajectory. Due to the complex non-convexity of the problem, we decompose the original problem into three sub-problems based on the alternating optimization (AO) structure. To be specific, we first derive the IRS reflection phase shift by using the semidefinite relaxation (SDR) and concave-convex procedure (CCCP) methods. Then, we obtain the computation and communication resource allocation by using the CCCP method. Finally, we determine the optimal AAV trajectory and decoding order by using the successive convex approximation (SCA) approach and the penalty-based method. A joint optimization algorithm is proposed for the AAV-MEC-NOMA framework. Numerical results demonstrate that: 1) The proposed algorithm is efficient and superior to other benchmark schemes in terms of increasing computational capacity;2) The application of IRS significantly enhances the computation capability of the network;3) The NOMA scheme outperforms the orthogonal multiple access (OMA) scheme.

关键词： Containers   Processor scheduling   Image edge detection   Scheduling   Collaboration   Performance evaluation   Servers   Computer science   Bandwidth   Resource management   Container scheduling   hierarchical structure of image   image layer reuse   mobile edge computing  

摘要： In Mobile Edge Computing (MEC) scenarios, computational tasks are popularly deployed using containerization to isolate the runtime environment. To complete the execution of the task, the edge server first pulls the image, then instantiates and runs the container. Since it takes a lot of time for the edge server to download the image from the cloud, image reuse reduces the pulling latency significantly. However, the limited storage capacity of edge servers hinders image reuse. Recent works have enhanced reuse efficiency by leveraging the hierarchical structure of images and caching high-value layers. However, their efficiency remains limited due to the lack of multi-container collaboration. This paper proposes a novel container scheduling strategy based on image layer reuse and sequence arrangement (ILR-SA) for MEC scenarios, which achieves efficient scheduling by collaborating multiple containers. First, containers are greedily deployed into the edge cluster. Then, the execution sequence of containers is modeled as an optimal Hamiltonian path problem, efficiently solved by our proposed decomposition algorithm. Finally, an efficient image layer update strategy is used to achieve layer reuse. We conduct rigorous experiments to demonstrate that our proposed container scheduling strategy reduces the computational task completion time by up to 91.3% compared to existing approaches.

关键词： Edge computing   Artificial intelligence   Training   Large language models   Cognition   Intelligent systems   Computer architecture   Computational modeling   Servers   Image edge detection   Edge AI   Mobile edge computing   edge general intelligence   large language models   small language models  

摘要： Edge Intelligence (EI) has been instrumental in delivering real-time, localized services by leveraging the computational capabilities of edge networks. The integration of Large Language Models (LLMs) empowers EI to evolve into the next stage: Edge General Intelligence (EGI), enabling more adaptive and versatile applications that require advanced understanding and reasoning capabilities. However, systematic exploration in this area remains insufficient. This survey delineates the distinctions between EGI and traditional EI, categorizing LLM-empowered EGI into three conceptual systems: centralized, hybrid, and decentralized. For each system, we detail the framework designs and review existing implementations. Furthermore, we evaluate the performance and throughput of various Small Language Models (SLMs) that are more suitable for deployment on edge devices. This survey provides researchers with a comprehensive vision of EGI, offering insights into its vast potential and establishing a foundation for future advancements in this rapidly evolving field.

关键词： Intelligent malware detection   Extreme learning machine   Edge computing environment   Independent component analysis   Hybrid wrapper with filter   Feature selection   Security threats  

摘要： Malware continues to wreak havoc on global digital ecosystems, with companies facing an average financial loss of $4.35 million per data breach in recent years. At the same time, individual users suffer from identity theft, affecting over 1.1 billion personal records annually. Existing malware detection systems often struggle with high latency in centralized cloud environments and fail to generalize across diverse malware variants generated by edge devices. To address these challenges, this work introduces MalwareNet, a novel multiclass malware detection network designed specifically for edge computing environments. MalwareNet innovatively processes data directly on edge devices, enabling real-time detection and classification with minimal latency and enhanced data privacy. The system employs a robust preprocessing pipeline to clean raw data, followed by Independent Component Analysis (ICA) to extract discriminative features while reducing dataset dimensionality. A Hybrid Wrapper-Filter (HWF) feature selection method optimizes feature subsets by integrating wrapper and filter techniques, ensuring compatibility with the chosen machine-learning classifier to maximize classification accuracy. The Extreme Learning Machine (ELM), selected for its rapid training and strong generalization, classifies malware into distinct categories, effectively identifying threats in edge settings. By combining edge-based processing, advanced feature engineering, and efficient classification, MalwareNet offers a scalable and reliable solution, significantly advancing malware detection capabilities for resource-constrained environments and providing a foundation for future adaptive security systems. Experimental evaluations on a large-scale malware dataset demonstrate the effectiveness of the proposed approach with an accuracy of 99.7 %, and F-measure of 99.55 %. The system also achieves high Jaccard index with an increment of 2.63 % in detecting and classifying malware, providing reliable

关键词： Costs   Array signal processing   Optimization   Resource management   Energy efficiency   Quality of service   Multiuser detection   Interference   Servers   Reconfigurable intelligent surface   cell-free networks   mobile edge computing   deep reinforcement learning   deep reinforcement learning  

摘要： This paper investigates the computation offloading problem in a distributed reconfigurable intelligent surface (RIS)-aided cell-free network, where users offload computationally intensive tasks to their associated access points (APs) with the assistance of multiple RISs. To minimize the long-term energy cost under latency constraints, we formulate a non-convex optimization problem aimed at jointly optimizing the offloading strategy, user transmit power, RIS reflection matrix, AP beamforming, and CPU frequencies at both users and the edge server. Given the problem's complexity and the interdependence of variables, we propose a two-stage decentralized proximal policy optimization (DPPO) algorithm to effectively manage the high-dimensional decision space. Simulation results demonstrate that our algorithm reduces the long-term energy cost by approximately 17% compared to existing methods, while ensuring that tasks are completed within the latency constraints.

关键词： Intelligent wireless control networks   Mobile edge computing   Multi-agent deep reinforcement learning   Dynamic multi-channel access   Traffic prediction  

摘要： Effective service distribution management is essential in Intelligent Wireless Communication Networks to meet the increasing Quality of Service (QoS) demands across various applications. Traditional transmission strategies often prioritize high-QoS data, which can lead to access starvation for lower-priority data in resource-constrained environments. To address this, we propose a QoS-aware adaptive service distribution strategy that balances the needs of high-and low-priority data without compromising the performance of either. Leveraging enhanced Multi-Agent Deep Deterministic Policy Gradient (e-MADDPG), our solution dynamically optimizes service distribution in mobile edge networks. By employing a gated recurrent unit-enhanced reinforcement learning framework, we enable intelligent agents to collaboratively decide channel access based on real-time traffic conditions. The proposed multi-criteria Decision-based Multi-channel Access algorithm allows high-priority data to defer access if necessary, improving the completion rates of lower-priority data. Furthermore, our method integrates network slicing and computation offloading to enhance service adaptability, ensuring efficient use of edge resources. Simulation results confirm that our framework significantly outperforms existing approaches in terms of channel utilization, QoS adherence, and overall network efficiency.