关键词： Mobile edge computing   Unmanned aerial vehicle   Task scheduling   Performance measures   Optimization algorithm  

摘要： This paper studies the task scheduling problem in the multiple unmanned aerial vehicle (UAV) assisted mobile edge computing (MEC) system. In order to ensure low delay and high energy utilization, a multi-layer heterogeneous task scheduling strategy is proposed. The strategy allows system model to simulate arrival and service of tasks by applying heterogeneous M/M/1, M/M/c, and M/M/c/K classical queues. In order to better evaluate the proposed strategy, we obtain various performance measures. By trading off delay and energy consumption, a single-objective optimization problem with constraints is formulated. Then we introduce the penalty function method to transform the problem into a single-objective optimization problem without any constraints, and propose a hierarchical heterogeneous task scheduling optimization (HHTSO) algorithm to optimize the formulated problem. Theoretical analysis and simulation results show that the proposed algorithm is more effective in ensuring low delay and low energy consumption.

关键词： Game theory   Nash equilibrium   Multiple video streams   Collaborative optimization   Adaptive offloading   Mobile edge computing   Bandwidth resources  

摘要： Massive video stream transmission and analysis require a large amount of bandwidth and computing resources, which poses a serious challenge to the current video stream offloading scheme based on mobile edge computing (MEC). A self-adaptive offloading scheme based on a balanced game multi video stream collaborative optimization framework is proposed for this purpose. Firstly, under the constraint of long-term MEC energy budget, the processing cost of video tasks is minimized by jointly optimizing data stream selection decisions, server offloading decisions, bandwidth resource allocation, and computing resource allocation. Secondly, an adaptive task offloading algorithm based on game theory and Nash equilibrium is designed to achieve optimal node selection through offloading algorithm and weight allocation, minimizing the processing cost of the entire multi video stream system and balancing video stream computation latency and energy consumption;Finally, the simulation results show that the proposed scheme satisfies long-term MEC energy constraints while significantly outperforming existing benchmark research schemes in terms of cost performance.

关键词： Mobile edge computing   Network Function Virtualization   Service Function Chain   Network Function Parallelism   Fairness-aware optimization   Deep Reinforcement Learning  

摘要： In Mobile Edge Computing (MEC), efficient resource allocation and optimal Service Function Chain (SFC) assembly are critical challenges in Network Function Virtualization (NFV). Parallelizing network functions can mitigate the high latency caused by the increasing number of Virtual Network Functions (VNFs). However, existing approaches often assume static resource demands during SFC assembly, which may not reflect real-world conditions. To address this limitation, this paper presents a Deep Reinforcement Learning (DRL)-based framework that dynamically optimizes network function parallelism while considering resource demand fluctuations. The proposed strategy models the problem as a fairness-aware throughput maximization task, ensuring balanced resource allocation and improved service efficiency. By leveraging a learning-based strategy, our strategy adapts to dynamic network conditions, reducing service deployment latencies and enhancing overall performance. Extensive simulations demonstrate that our strategy outperforms the best state-of-the-art method in terms of average latency and deployment rate, achieving improvements of 4.3% and 1.98%, respectively.

关键词： Mobile edge computing   Cache replacement   UAV cooperation   Federated learning   Reinforcement learning   Age of information  

摘要： Existing mobile edge computing caching strategies face several challenges, including privacy leakage due to centralized training, the lack of consideration for the age of cached content, and issues such as poor content freshness, low cache hit rate, and high system energy consumption, which arise from the randomness and high dynamics of user requests in emergency scenarios. To address these issues, a federated reinforcement learning-based multi-unmanned aerial vehicle cooperative dynamic caching replacement strategy is proposed. Firstly, a comprehensive utility objective function is developed, considering user access delay, the age of cached content, and system energy consumption. Secondly, BiLSTM is employed to capture the dynamic dependencies in user request time series, combined with an attention mechanism to focus on key features. Federated learning is used to achieve distributed training across UAVs, ensuring user privacy while improving the accuracy of content popularity prediction. Subsequently, a weighted caching scoring mechanism based on content popularity and information age is proposed to dynamically adjust the cache replacement order. Finally, the cache replacement and user association problems are modeled as a Markov decision process, and the TD3 reinforcement learning algorithm is applied, utilizing double Q-networks and a delayed policy update mechanism, to derive the optimal cache replacement strategy. The experimental results demonstrate that the proposed method outperforms comparison approaches in terms of cache hit rate, access delay, and content freshness, demonstrating strong performance advantages and application potential.

关键词： Cloud computing   Delay Awareness   Edge computing   Mobile Edge computing   Queuing theory   Resource Allocation   Task scheduling  

摘要： The key aspect of Mobile Edge Computing (MEC) that has aroused considerable concern is how to efficiently schedule tasks and allocate resources to meet real-time applications' low latency and energy requirements. Classical scheduling algorithms such as First-Come-First-Served (FCFS), Shortest Job First (SJF), and Round Robin (RR) are largely ineffective for managing dynamic workloads and heterogeneous resource environments. This paper proposes A delay energy-efficient queuing-based Particle Swarm Optimization (DAEE-QPSO) algorithm that combines queuing theory and swarm intelligence to optimize task offloading decisions. Individual nodes of edges are represented by M/M/1 queues that allow the queuing delay and system congestion to be estimated perfectly. A hybrid fitness function is constructed that attempts to concurrently reduce the runtime and energy usage of tasks and constraints (meeting task deadlines and energy budgets). The proposed DAEE-QPSO algorithm shows the lowest average run time and energy consumption compared to FCFS and EADF-PSO. It also has high task throughput and maintains the greatest average energy saving of up to 29.8%, outperforming Energy Aware Double Fitness Function PSO (EADF-PSO) by only 23.3%. This result demonstrates its effectiveness in improving task efficiency in a mobile edge cloud. The suggested architecture is very applicable to energy-conscious and real-time usage in the cloud-to-edge environment. © 2025 Seventh Sense Research Group®.

关键词： Deep learning   DNN model partitioning   mobile edge computing   mobile vision application   of experience   quality   resolution control  

摘要： With the advancement of mobile device and network performance, the industry for mobile vision applications, such as image classification and object detection, is rapidly growing. When designing mobile vision applications, employing mobile edge computing (MEC) based on deep learning model partitioning can significantly improve the quality of user experience (QoE). Research on deep learning model partitioning has advanced in various directions over time, but until now, there has been no study that simultaneously partitionizes deep learning models by layers and adjusts the input frame size. Our proposed algorithm, Parecon, is based on Lyapunov optimization and dynamically adjusts 1) the number of frames to be processed, 2) model partition points, and 3) input frame size at each time slot. The proposed algorithm simultaneously optimizes processed fps, E2E latency, and top-1 accuracy. Through simulations, we confirmed that Parecon achieves a significant improvement in fps compared to existing algorithms while maintaining simular E2E latency and top-1 accuracy. © 2025, Korean Institute of Communications and Information Sciences. All rights reserved.

关键词： Resource management   Optimization   Servers   Vectors   Fading channels   Data communication   Scalability   Massive MIMO   Convolutional neural networks   Artificial intelligence   Cell-free   CNN   ICA   latency   mobile edge computing   resource allocation  

摘要： This letter investigates the latency minimization at the network edge in mobile edge computing (MEC)-enabled Cell-Free massive MIMO systems. We introduce a new edge computing model that integrates both task offloading and local execution. To minimize overall system latency while considering power allocation constraints, we formulate an optimization problem aimed at reducing maximum computing time. This mixed-integer non-convex problem is then reformulated into a more tractable form, which is solved using an iterative convex approximation method to achieve locally-optimal solutions. Additionally, we propose a convolutional neural network-based algorithm as an alternative solution to further improve system efficiency. Numerical results are provided to validate the theoretical framework and demonstrate the effectiveness of the proposed approaches in accelerating the data processing in MEC-enabled cell-free networks.

关键词： Processor scheduling   Servers   Costs   Resource management   Pricing   Indexes   Convergence   Adaptive scheduling   Training   Decision making   Mobile edge computing   resource scheduling   adaptive timescales   reinforcement learning  

摘要： In mobile edge computing (MEC), resource scheduling is crucial to task requests' performance and service providers' cost, involving multi-layer heterogeneous scheduling decisions. Existing MEC schedulers typically adopt static-timescale scheduling, where scheduling decisions are updated regularly at fixed intervals for all layers. The inflexible updating timescales lead to poor performance in the production networks. In this paper, we propose EdgeTimer, an unprecedented approach that automatically and adaptively determines respective updating timescales of multiple scheduling layers to achieve a better trade-off between the operation cost and service performance. Specifically, we design (i) a three-layer hierarchical deep reinforcement learning (DRL) framework for efficient learning of tightly coupled policies, (ii) a tailored multi-agent DRL algorithm for decentralized scheduling, with the convergence strictly proved, and (iii) a lightweight system defender for deterministic reliability assurance. Furthermore, we apply EdgeTimer to a wide range of Kubernetes scheduling rules, and evaluate it using production traces with different workload patterns. Through extensive trace-driven experiments, we demonstrate that EdgeTimer can significantly decrease the operation cost for service providers without sacrificing the delay performance, thereby improving overall profits, compared with the state-of-the-art approaches.

关键词： Mobile edge computing   movable antenna   discrete antenna position selection   discrete antenna position selection   resource allocation   resource allocation   resource allocation  

摘要： In this work, movable antenna (MA) is invoked for enhancing the task completion performance of a mobile edge computing (MEC) system. Specifically, in the considered system model, a base station equipped with MAs assists in computing tasks offloaded from multiple devices. In contrast to previous works focusing on continuous antenna movement, a more realistic scenario involving discrete antenna position selection is taken into account under the constraint of minimum antenna distance. To maximize the total task completion rate of all devices, it is essential to jointly optimize the transmit power and computation rate at each device as well as MA position selection while satisfying power and discrete antenna selection constraints. To address this challenging mixed-integer nonlinear programming problem, we first reformulate the original problem into an equivalent, more tractable form by Lagrangian dual transform. Then a penalty dual decomposition based method is proposed to efficiently find the stationary solutions. Simulation results confirm that MA-assisted MEC systems significantly outperform the conventional fixed-position antenna and antenna selection based schemes.

关键词： Mobile edge computing   internet of things   deep reinforcement learning   communication security   task offloading optimization  

摘要： As the network society progresses, Mobile Edge Computing (MEC) has emerged as a transformative technology, enabling low-latency and high-efficiency data processing, which is reshaping the evolution of the Internet of Things (IoT). By offloading computational tasks to the network edge and delivering services closer to the data source, MEC effectively mitigates data transmission latency and enhances system responsiveness, positioning itself as a pivotal enabler of future intelligent societies. This study addresses the critical issue of communication security in urban environments, where eavesdropping poses significant threats, by proposing a novel joint optimization framework that integrates UAV flight trajectory planning, user scheduling and computational resource allocation. To navigate the complexities of high-dimensional state and continuous action spaces, we leverage the Deep Deterministic Policy Gradient (DDPG) algorithm for iterative policy optimization. The objective is to minimize task processing delays while ensuring secure communication. Experimental results substantiate the efficacy of the proposed algorithm, demonstrating its ability to achieve a balanced trade-off between communication security and task processing efficiency, across varying weight coefficients, thereby enhancing the overall performance of MEC systems in IoT contexts.