关键词： Mobile edge computing   end-to-end delay   virtual reality   neural frame interpolation  

摘要： In this article, with the objective of significantly increasing the frame rate of virtual reality (VR) videos, we design an efficient end-edge collaborative VR streaming system which consists of three modules: frame similarity analysis, offloading decision making, and collaborative frame interpolation. In specific, frame similarity analysis tries to eliminate redundant frames based on perceived quality assessment, so that the required number of interpolated frames can be reduced without deteriorating visual quality. Then, an end-to-end (E2E) delay optimization problem is formulated to obtain the optimal offloading strategy, by balancing the transmission and computing burden of neural frame interpolation via end-edge collaboration. Furthermore, the E2E delay of the proposed system is theoretically analyzed based on queuing theory. Our analysis reveals that, the proposed collaborative distribution of interpolation tasks between edge and end devices are effective to achieve the minimal E2E delay of streaming VR videos. Extensive experimental results demonstrate that the proposed system can significantly improve the frame rate of VR videos, while maintaining timely VR content delivery in various networking conditions.

关键词： High-speed mobile business   mobile edge computing   service component caching   computation task offloading  

摘要： With the rapid development of high-speed assistant driving and smart inspections, the edge network is required to provide quick and reliable service to avoid large service response delays and frequent re-transmissions caused by interruption. However, the reasonable service component caching, efficient edge collaboration and reliable cross-domain computation offloading are still key problems to be solved. Thus, we consider an AI and mobile edge computing (MEC) integrated service framework, which is highly reliable for high-speed mobile businesses, and we divide the service process into component caching phase and task offloading phase. In the first phase, we novelly define the edge collaborative service domain (ECSD) which allows multiple edge nodes to collaboratively share resources from a global perspective and design a user behavior aware service component pre-caching method to increase resource utilization. In the second phase, based on the formed ECSDs and cached service components, we present an AI-empowered cross-domain computation task offloading mechanism including task partition and backup to enhance the reliable service capability of edge networks. Simulation results verify that the proposed mechanism can jointly optimize the allocation of caching, computation, and communication resources, while improving the service response speed and resource utility of edge networks.

关键词： Federated learning   intermediate result exchanging   mobile edge computing   P2P communication  

摘要： The emerging Federated Learning (FL) permits all workers (e.g., mobile devices) to cooperatively train a model using their local data at the network edge. In order to avoid the possible bottleneck of conventional parameter server architecture, the decentralized federated learning (DFL) is developed on the peer-to-peer (P2P) communication. In DFL, model exchanging among workers is usually regarded as an atomic operation, which largely affects the total bandwidth consumption during model training. Given the limited communication resource on workers, model exchanging will pose a great challenge when meeting with the large-scale models. Herein, we propose to let workers exchange the intermediate results, instead of the entire model, with each other. We provide theoretical analysis of DFL based on intermediate result exchanging, which reveals the relationship between the training performance and the exchanging interval (i.e., the number of local updating iterations) of intermediate results. According to the convergence bound, we propose an adaptive exchanging interval (or frequency) algorithm called Fed-IR, which optimizes the trade-off between communication cost and training performance. Extensive simulation results show that compared with the model exchanging methods, our proposed algorithms can save communication traffic of around 42%similar to 81% while still achieving the similar accuracy.

关键词： Mobile edge computing   Emergency tasks   SDN   Collaborative computation offloading  

摘要： Due to the growing demand for time-intensive tasks, several tasks coexist, some of which last for a long time or only occur once in a while, such as emergencies. Emergencies such as natural disasters, accidents or medical emergencies require immediate attention and often demand real-time execution. Such tasks are the most difficult to plan for and foresee. Planning, resource allocation, and clear communication between all involved parties and network layers are crucial for scheduling emergency tasks. In the extensive literature, when an emergency task triggers, it will be handled like any other regular task. If the deadline is missed, a catastrophe will ensue. They ignored the impact of insufficient mobile edge computing (MEC) capacity and network congestion, which might result in execution delays. Collaborative offloading, which involves distributing computation tasks across multiple servers, shows promise in enhancing MEC network performance, particularly in emergency scenarios. In this context, efficient task scheduling plays a vital role in minimizing the total execution time of regular tasks while meeting the deadlines of emergency tasks. Our proposed scheme utilizes a collaborative offloading approach to schedule emergency tasks in MEC networks, leveraging the computing capacity of edge-deployed MEC servers. By utilizing the software-defined networking (SDN)'s global view of the network, task requests are collected and allocated to suitable MEC servers capable of meeting the demands. To address key challenges, our scheme propose four scheduling algorithms to address the following issues: (i) ensuring tasks are assigned to the nearest MEC server with sufficient computational resources, (ii) controlling a threshold to prevent network congestion, (iii) selecting an optimal collaborative MEC server for executing overloaded tasks based on collaborative offloading decisions, and (iv) allocating resources for emergency tasks when triggered to meet urgent deadlines by

关键词： Mobile edge computing  

摘要： Mobile edge computing (MEC) is increasingly being applied to task computation with the continuing increase of computing-intensive applications in vehicular networks. The MEC can offload the tasks generated by the vehicle to the MEC server to mitigate the computational constraints on the mobile vehicle. However, most task offloading algorithms can not cope with the dynamic distribution of vehicles in vehicular networks. Moreover, most task offloading algorithms do not consider vehicular networks using multipath transmission schemes. Therefore, it remains a challenge to implement efficient joint task offloading and resource allocation (JTORA) algorithms in multipath transmission vehicular networks. In this paper, we introduce a novel cooperative transmission network architecture in multipath transmission vehicular networks. The computational tasks generated by the vehicles can be offloaded to MEC servers deployed near the base station or to cloud servers via multipath transmission. We formulate the task offloading and resource allocation problem as a multi-objective optimization problem in a multipath transmission vehicular network. To achieve efficient task offloading and resource allocation under cooperative transmission network architecture, we propose a JTORA algorithm for multipath transmission vehicular networks based on deep reinforcement learning (DRL). JTORA converts the optimization problem into a DRL model and obtains a specific optimization policy through model training. Simulation results show that JTORA achieves lower average task delay and lower average task cost than other traditional heuristic algorithms. In this article, a joint task offloading and resource allocation (JTORA) algorithm is proposed for multipath transmission vehicular networks based on deep reinforcement learning (DRL). Simulation results show that JTORA achieves lower average task delay and lower average task cost than other traditional heuristic ***

关键词： Mobile edge computing   Truthfulness   Resource allocation   Energy consumption   Algorithm design  

摘要： This paper addresses multi-access multi-server multi-task resource allocation in mobile edge computing (MEC). Our aim is to maximize social welfare for heterogeneous MEC servers from multiple access points (APs) providing heterogeneous virtual machine instances to the mobile devices (MDs) within the coverage area. In the system model, each MD has multiple tasks, and tasks can offload to more than one MEC server through different APs within its direct communication range. We formulated the problem with an auction-based model to provide flexible service. However, the MDs are self-interested and can misreport their preferences, leading to inefficient resource allocation. We designed an optimal approximation mechanism to solve this problem. Then, we showed that this mechanism achieves individual rationality and truthfulness, that is, the MDs had no incentive to declare untrue values. In addition, we analyzed the approximation ratio of our truthfulness mechanism. The task allocation problem was also considered, and the proposed approximation algorithm could stop at any step and provide reasonable performance. Experimental results demonstrated that our proposed approximation mechanism provides near-optimal social welfare in a reasonable time and effectively reduces energy consumption.

关键词： Mobile edge computing   load balance   mobility-aware task offloading   reinforcement learning  

摘要： Mobile edge computing is a promising computing paradigm enabling mobile devices to offload computation-intensive tasks to nearby edge servers. However, within small-cell networks, the user mobilities can result in uneven spatio-temporal loads, which have not been well studied by considering adaptive load balancing, thus limiting the system performance. Motivated by the data analytics and observations on a real-world user association dataset in a large-scale WiFi system, in this paper, we investigate the mobility-aware online task offloading problem with adaptive load balancing to minimize the total computation costs. However, the problem is intractable directly without prior knowledge of future user mobility behaviors and spatio-temporal computation loads of edge servers. To tackle this challenge, we transform and decompose the original task offloading optimization problem into two sub-problems, i.e., task offloading control (ToC) and server grouping (SeG). Then, we devise an online control scheme, named MOTO (i.e., Mobility-aware Online Task Offloading), which consists of two components, i.e., Long Short Term Memory based algorithm and Dueling Double DQN based algorithm, to efficiently solve the ToC and SeG sub-problems, respectively. Extensive trace-driven experiments are carried out and the results demonstrate the effectiveness of MOTO in reducing computational costs of mobile devices and achieving load balancing when compared to the state-of-the-art benchmarks.

关键词： Augmented reality   deep reinforcement learning   mobile edge computing   task offloading  

摘要： With the popularity of mobile devices and the continuous advancement of mobile network technology, running online augmented reality (AR) on lightweight mobile devices is much more desirable than on heavy and expensive head-mounted devices that are difficult to satisfy users. Mobile edge computing can assist in supporting AR applications running on mobile devices, which copes with compute-intensive and delay-sensitive requirements. However, subject to the limited and heterogeneous edge resources, offloading tasks to edge devices is not easy, especially if the application requires multi-party interaction. It is challenging to develop a credible task placement scheme that satisfies user experience with flexible use of edge resources. This article focus on the task offloading placement problem for AR overlay rendering in multi-party mobile augmented reality system. We first present our observations about performance bottlenecks of edge devices and explain the necessity of splitting the AR overlay rendering pipeline. We then formulate a joint optimization problem of task placement decisions, aiming to maximize the user experience of quality and minimize the service cost. We develop a novel decision approach based on deep reinforcement learning (DRL) to address this complex problem. Finally, we verify the effectiveness and superiority of the proposed method through extensive evaluation experiments.

关键词： Task analysis   Games   Resource management   Multitasking   Bandwidth   Wireless communication   Costs   Multi-task   mobile edge computing   computation offloading   potential games   Nash equilibrium  

摘要： Mobile edge computing emerges to serve mobile users with low-latency computation offloading in edge networks, which are resource-constrained with massive users and workloads. However, existing communication and computing resource allocation schemes for offloaded tasks aren't efficient enough, where finished tasks still occupy resources, wasting constrained resources. Besides, the multi-user offloading is usually for scenarios of one task per user, ignoring real-world multi-task offloading scenarios where each user has multiple tasks, lack generality and flexibility. Meanwhile, local computing resource allocation schemes in multi-task scenarios ignore resource readjustment, causing low resource utilization. To solve these problems, we propose ECO-GAME, an efficient multi-task offloading scheme, which dynamically allocates bandwidth and computing resources to unfinished tasks, resulting in high resource utilization. We initially formulate the multi-task offloading problem as the game minimizing each user's cost, which is NP-hard. Thus we re-formulate the game utilizing potential games to optimize user's objective either locally or globally, and prove the existence of its Nash equilibrium. We then design an efficient multi-task offloading algorithm to obtain an approximate solution in polynomial time, together with computational complexity analysis. We further conduct performance evaluation on ECO-GAME utilizing price of anarchy. Numerical results demonstrate the efficiency of ECO-GAME, and show ECO-GAME reduces 49.2% cost over the state-of-the-art work, and scales well with the increasing number of tasks and users.

关键词： Industry 5.0   Mobile edge computing   Joint optimization   Computational energy efficiency   Partial offloading  

摘要： Smart devices in Industry 5.0, such as sensors and robots, are often limited by low battery life and finite computational resources, hindering their ability to perform complex tasks. By offloading computation-intensive tasks to Mobile Edge Cloud Computing (MEC) servers at the network's edge, businesses can achieve real-time data processing and analysis, reducing communication latency, quicker response times, and improved system reliability. This work presents an integrated framework for MEC and Industry 5.0, aimed at enhancing the performance, efficiency, and flexibility of industrial processes. In particular, we propose a joint optimization problem that maximizes computational energy efficiency by optimally allocating resources, such as processing power and computational resources, as well as device association, in the most efficient manner possible. The problem is formulated as nonconvex/nonlinear, which is intractable and poses high complexity. To solve this challenging problem, we first transform and decouple the original optimization problem into a series of subproblems using the block coordinate descent method. Then, we iteratively obtain an efficient solution using convex optimization methods. In addition, our work sheds light on the fundamental trade-off between local computation and partial offloading schemes. The results show that for small data size requirements, the performance is comparable among different schemes. However, as data size increases, our proposed hybrid scheme, which includes a partial offloading scheme, outperforms others, highlighting the effectiveness of the proposed joint optimization scheme.