关键词： Federated learning   Computational modeling   Privacy   Data models   Training   Differential privacy   Multi-access edge computing   mobile edge computing   consumer electronics   deep reinforcement learning   local differential privacy  

摘要： Consumer electronic devices often involve processing and analyzing a large amount of user personal data. Nevertheless, owing to apprehensions regarding privacy and security, users are hesitant to transmit this sensitive data to centralized cloud servers for training. The combination of mobile edge computing and federated learning (FL) enables local devices to access computational power and storage resources, allowing them to engage in distributed learning and model training while safeguarding user privacy. However, these resources are not unlimited. Furthermore, as artificial intelligence technology progresses, inference attacks have become a major threat to privacy in traditional federated learning. To address these challenges, we propose an innovative federated deep learning algorithm, called Fed-PEMC. This algorithm combines local differential privacy and model compression techniques. By leveraging deep reinforcement learning for model compression, Fed-PEMC reduces model size while maintaining model accuracy, improving communication efficiency. We also introduce customized label sampling to accelerate model training. Before uploading the model, we implement local differential privacy protection on the compressed model, reducing privacy budget and addressing privacy leakage caused by inference attacks. Theoretical analysis and experimental results validate that Fed-PEMC adheres to (epsilon, delta) -differential privacy and exhibits a communication cost linked to the model size. Experimental results show that compared to baseline algorithms, Fed-PEMC excels in ensuring privacy, maintaining model accuracy, and optimizing communication efficiency, and Fed-PEMC outperforms the baseline solution DP-Fed by 2.27 and 2.02 percentage points in testing accuracy on the Mnist and Cifar10 datasets, respectively.

关键词： Mobile edge computing   Real-time video analysis   Computation offloading  

摘要： Driven by the Quality of Experience (QoE) demands for video analysis applications within contexts such as smart cities, Industrial Internet of Things (IoT), and Internet of Vehicles (IoV), networks are confronted with numerous challenges, including, but not limited to, the requirements for ultra-high bandwidth, ultra-low latency, reliability, and stability. Mobile Edge Computing (MEC) as a technology aims to decentralize computational and storage resources closer to data sources and end-users, meeting the demands of computation-intensive and latency-sensitive video analysis applications and thereby effectively addressing QoE requirements. This has led to a new intersection, edge video offloading, which has attracted widespread attention. However, to date, there are but a few extant papers on this topic with limited relevance. To bridge this gap, we offer a comprehensive review of the state-of-the-art approaches in computational offloading techniques applied to video data, innovatively revisiting existing task offloading technologies from the perspective of mobility of edge server nodes, including solutions based on both fixed and dynamic MEC servers. Additionally, we explore the differences in video offloading techniques between scenarios involving a single MEC server and those involving multiple MEC servers. Meanwhile, open challenges are presented to direct future research in this field.

关键词： Mobile edge computing   Mobility prediction   Relay assistance   Computing offloading   Unequal splitting of tasks  

摘要： Internet of Vehicles (IoV) is paving the road for the new generation of Intelligent Transportation Systems (ITS), and Mobile Edge Computing (MEC) is enabling IoV to efficiently handle the computation -intensive and time -sensitive tasks. However, this has introduced new challenges such as maximizing computing resources, allocating resources fairly for multi -source tasks concurrently, and dividing tasks for parallelly processing to minimize the latency. To face these challenges, a three-dimensional road vehicle mobility model is constructed, and the problem of offloading strategy and resource allocation among multiple vehicles served by one Road Side Unit (RSU) is investigates to minimize the average latency of multi -source tasks while satisfying the quality of service requirements. To address the Non -deterministic Polynomial -time hardness (NP -hardness) of the problem, we design a Relay -Assisted Parallel Offloading (RAPO) strategy to obtain the optimization solution. Extensive experimental results show that the RAPO strategy introducing relay -assisted nodes can enhance performance in poor scenarios and ensure low -latency multi -tasking under various conditions, especially reducing latency by 39% compared to local computing.

关键词： Task analysis   Servers   Autonomous aerial vehicles   Relays   Computational modeling   Resource management   Optimization   Mobile edge computing   unmanned aerial vehicle   relaying   user mobility   3D deployment  

摘要： As one of the most popular consumer electronics, Unmanned Aerial Vehicle (UAV) has the potential to assist User Equipment (UE) in the Internet of Everything. Mobile edge computing (MEC) is also an emerging technique that can provide sufficient computing resources to IoE users. In this paper, we focused on the UAV-assisted MEC problem with mobile UEs, where the UAV serves as an MEC server to assist UEs in computing and acts as a relay to deliver tasks to a ground access point (AP). The objective is to minimize the average time for completing all tasks in the network while jointly optimizing resource allocation such as communication bandwidth, CPU frequency, task division ratio, and UAV's three-dimensional location deployment. First, we proposed an online MEC network adjustment scheme. Then, we decomposed the formulated non-convex problem into three low-complexity subproblems. Last, we proposed a successive convex approximation-based joint optimization algorithm to solve them. In addition, we presented heuristic conclusions in task allocation rules, UAV flight trajectory prediction, and UAV hovering altitude selection, which can further reduce task completion time and deepen the understanding of the working mode of mobile edge servers. Simulation results show that the algorithm can significantly reduce the completion time.

关键词： Federated learning   Training   Computational modeling   Data models   Servers   Error compensation   Quantization (signal)   Mobile edge computing   federated learning   machine learning   gradient compression   communication efficiency  

摘要： In this paper, we explore an emerging customer application named federated learning in Mobile Edge Computing (MEC). Federated learning is a machine learning technique that enables data privacy and avoids data silos. However, federated learning tasks with high performance often face large model parameters and high communication costs. To address this challenge, this paper presents FedSQ (Federated Learning with Sparsity and Quantization), a novel sparsity and quantization strategy that reduces communication overhead and enhances model convergence. FedSQ selectively uploads the parameters with large gradient change ratios and uses the old values from the previous round for the parameters with small gradient change ratios. FedSQ also incorporates an error compensation mechanism with correction to mitigate the compression loss and balance the model parameters between the current and previous rounds. We conduct experiments on standard datasets and demonstrate that FedSQ achieves high compression ratio without sacrificing the model accuracy significantly in various federated learning tasks. Across all communication rounds, FedSQ achieved the highest accuracy improvement of 25.74% and the highest compression rate of 97.98%. Our research provides a novel intelligent data-driven model for consumer electronics, which can improve the communication efficiency and model convergence of client applications in MEC.

关键词： Task analysis   Servers   Resource management   Processor scheduling   Consumer electronics   Interference   Wireless communication   Mobile edge computing   online   deadline-sensitive   resource allocation   competitive ratio  

摘要： With the remarkable prosperity in smart consumer electronics (CEs), Mobile Edge Computing (MEC) has emerged as a pivotal technology to tackle the prevalent latency issues faced by smart CEs. Consequently, this has produced multiple deadline-sensitive tasks, imposing stringent computational time requirements on the infrastructure of 6G wireless communication systems. To meet the deadline demands of tasks when offloading, research efforts have focused on improving the scheduling mode. However, a majority of modes are offline, which is unrealistic for MEC servers to schedule tasks or reserve resources according to the global information (i.e., the time and quantity of tasks released by CEs) grasped in advance. To address this concern, we propose an online data-driven scheduling mechanism maximizing the revenue of deadline-sensitive tasks in the MEC-enabled consumer electronics system. Given that the released time and deadline time are fixed, but the execution process are flexible and the execution process involves two network resources, we design a two-step online resource allocation (TORA) algorithm comprising an online spectrum allocation sub-algorithm and an online computing resource allocation sub-algorithm. Moreover, we derive a precise competitive ratio to evaluate the performance of our TORA algorithm. Finally, through extensive simulations, we demonstrate its superiority in improving system revenue.

关键词： Mobile edge computing   Edge intelligence   Service placement   Deep deterministic policy gradient  

摘要： Mobile edge computing is an energy-efficient paradigm which has great support for resource-limited user devices performing compute-intensive programs and applications. However, limited edge resources, mobility of user equipment, growth of service requirements, and dynamic nature of service types make it a challenging task to configure computing and storage resources for executing various services on edge servers. Therefore, an adaptive joint service placement framework in the edge system with various devices of user and mobile edge servers is proposed in this paper. The proposed framework takes into account the mobility of user devices, the dynamic changes in various types of services, the cost of service placement and service usage, and optimizes the service placement scheme from the perspective of different target groups. Simultaneously, we design a deep deterministic policy gradient based service placement tuning approach in which centralized critic networks and actor networks are jointly used to improve the service placement performance. The relative evaluation results validate the effectiveness of the proposed framework and approach in improving the performance of the edge system.

关键词： mobile edge computing   computing offloading   serial and parallel priority   improved graphgravitational search algorithm   genetic algorithm  

摘要： Mobile edge computing (MEC) distributes computing and storage resources to the edge of the network closer to the user and significantly reduces user task completion latency and system energy consumption. This paper investigates the problem of computation offloading in a three-tier mobile edge computing network composed of multiple users, multiple edge servers, and a cloud server. In this network, each user's task can be divided into multiple subtasks with serial and parallel priority relationships existing among these subtasks. An optimization model is established with the objective of minimizing the total user delay and processor cost under constraints such as the available resources of users and servers and the interrelationships among the subtasks. An improved gravitational search algorithm (IGSA) is proposed to solve this optimization model. In contrast with the other gravitational search algorithm, the convergence factor is introduced in the calculation of the resultant force and the crossover operation in a genetic algorithm is performed when generating the new particles during each iteration. The simulation results show that the proposed IGSA greatly improves the system performance compared with the existing algorithms.

关键词： Mobile edge computing   Task scheduling   Energy harvesting   Renewable energy   Energy storage unit   Sample average approximation  

摘要： Aiming to provide low-cost MEC services for mobile devices in areas without/lacking infrastructure, we propose a renewable-energy-driven energy-harvesting-based task scheduling and energy management framework. First, for mobile devices, the constructed energy consumption minimization problem is solved by an alternating-optimization-based algorithm. Then, we design an energy management algorithm based on sampling average approximation to derive the optimal charging/discharging strategies, number of energy storage units, and renewable energy utilization. The simulation results show that the proposed framework can significantly reduce the cost of MEC servers and prolong the working hours of mobile devices. (c) 2023 The Author(s). Published by Elsevier B.V. on behalf of The Korean Institute of Communications and Information Sciences. This is an open access article under the CC BY license (http://***/licenses/by/4.0/).

关键词： Digital twins   Biological system modeling   Cloud computing   Solid modeling   Sensors   Resource management   Computational modeling   Customers centric network   mobile edge computing   digital twin   smart data-driven modeling   content distribution  

摘要： Recently, the method of non-intelligent vehicles, acting as information consumers, receiving integrated data shared to the cloud by mobile edge nodes and RSUs has garnered significant interest for its potential to enhance road safety. The method can assist the customers in achieving advanced driver assistance system functionalities. Nonetheless, challenges such as resource constraints at RSUs, long communication distances between the cloud and vehicles, and frequent data interactions make it difficult to meet the strict latency requirements in highly mobile and complex traffic environments. To address these challenges, this paper proposes a smart data driven multi-level synchronous digital twin model for vehicle-assisted driving. We make use of multi-level synchronous deduction along with high-level proactive updating to reduce update latency significantly so that it can meet the timeliness requirements for cyber-physical services in differentiated scenario. Additionally, we construct a customer matching value and dynamic vehicle model update system. It solves the resource allocation problem that arises during the update of differentiated twin models for multiple customers when resources of RSUs are constrained. While satisfying the constraints of update latency, it minimizes the cost of vehicle model updates and effectively enhance system efficiency.