基于联邦元学习的安全移动边缘计算卸载框架

移动边缘计算(MEC)技术通过卸载部分计算任务到边缘服务器,可将第5代网络(5G)、云计算、大数据和人工智能等技术延伸到物联网终端。针对如何高效卸载计算任务和保障边缘数据隐私安全2个关键问题,在综述计算卸载性能优化研究基础上,本文提出了一种融合联邦学习和元学习的计算卸载应用框架,通过对计算任务的计算卸载以及计算资源的联合优化,从而实现系统加权时延和最小。在不泄露用户数据隐私的前提下,联合多个边缘服务器共同训练一个全局模型,并实现边缘服务器个性化计算卸载应用。在新的计算任务场景下,全局模型的网络参数仅用少量训练样本就能快速收敛。实验测试结果表明,本文提出的基于联邦元学习的计算卸载框架可适应未来边缘计算应用的隐私安全需求。     

基于边缘计算与区块链的车载计算资源智能调度研究

随着车联网(IoV)的快速发展及部署,用户对网络服务质量的要求也随之提高。车联网数据计算作为网络服务的重要内容之一,越来越受到关注。移动边缘计算(MEC)作为一种允许车辆将计算任务卸载到车联网系统边缘服务器的技术,能够有效降低计算时延,提升数据处理效率。然而,车联网的数据流量日益增加,导致边缘计算设备的需求量大幅提高且存在数据安全可靠性问题。对此,本文面向车联网中移动车辆计算卸载的场景,提出一种基于区块链的停放车辆辅助计算的系统模型。通过联合考虑服务器计算资源、车辆机动性等条件,利用深度强化学习(DRL)对计算卸载和资源分配策略进行优化,减少系统能耗和数据传输时延,并提高区块链系统的交易吞吐量。仿真结果表明,本文所提优化方法可以有效提升系统性能,同时具有良好的收敛性能和稳定性。     

房间空调器国内外能效指标体系、测试及计算方法差异分析

详细介绍并对比分析了中国、美国、日本、印度、沙特及欧盟地区房间空调器的能效标准指标体系、测试方法和计算方法,以期读者能够了解和掌握房间空调器主要应用国家或地区的相关情况,为房间空调器产品设计及推广提供有效信息和帮助.     

移动边缘计算增强的超密集网络上行传输性能分析与基站配置设计

在引入移动边缘计算(MEC)技术的超密集网络(UDN)中,网络性能受无线传输链路质量和基站计算资源部署的共同影响。本文以小型基站接收干扰信号的统计特性分析为基础,对UDN与MEC结合场景的上行空间遍历容量进行分析,并根据空间业务强度和小基站计算服务排队系统的稳定性约束关系,进行小基站配置设计。首先,采用空间泊松点过程对干扰源用户分布进行建模,同时考虑无线信道的多天线与小尺度衰落特性,以小基站部署密度为变量分析上行用户信干比的统计特性及变化规律;然后,根据排队论计算空间业务强度与MEC服务器计算能力之间的约束关系;最后,采用数值仿真验证了信干比与基站密度关系分析的正确性,得出了空间遍历容量的收敛趋势,并给出了小基站与MEC服务器配置设计的例子。     

无线网络中基于深度强化学习的资源分配研究

由于用户设备的多样性,直播视频必须转换为不同的格式.且由于无线网络环境的动态特性,为用户提供高质量和严格的时延要求的实时视频是一个很大的挑战.本文提出了一种联合优化用户调度、视频质量选择和资源分配的方案,以达到在视频直播业务中最大化视频质量的同时尽可能减少播放延迟的目标.通过深度Actor-Critic强化学习算法进行仿真.仿真结果表明,本文提出的方案可以提高用户视频体验质量(QoE),并且相比于策略梯度算法(PG),本文算法学习速度更快.     

基于部分可观察马尔可夫决策过程的受控无线网络系统动态资源分配

研究了受控无线网络的动态资源分配。针对传统无线通信传输模型的局限性随着无线通信系统架构的发展日益凸显的问题,提出了一种引入反馈控制策略的受控无线网络模型。该模型结合部分可观察马尔可夫决策过程(POMDP),将用户接收功率与数据传输误码率作为反馈观测对象,对通信小区内基站天线开启数与用户接入数进行动态资源最优匹配。仿真结果表明,这种方法能够有效提升系统传输能效性与可靠性,降低传输误码率,改善系统资源动态匹配控制性能。     

身边的移动计算

身边的新技术发展之快,使用的新设备更新之快,令我们不得不感慨,当今时代发展之迅猛。本文队身边的移动计算为例从概念、技术、应用等方面浅谈了我们都能感受到的移动计算。     

基于边缘检测的神经网络非均匀性校正及硬件实现

红外焦平面探测器的非均匀性校正技术仍然是当前红外热成像系统重点研究的关键技术之一。相对定标类算法,基于场景非均匀性校正根据场景进行非均匀参数更新,不需要使用挡板遮挡视场。本文介绍传统神经网络非均匀性校正算法及加快收敛速度的改进措施,引入边缘检测方法来克服传统神经网络算法的鬼影问题。文中阐述的方法已在以TMS320DM643为处理核心DSP硬件处理平台上实现,取得了较好的校正效果。     

雾计算辅助车联网中面向视频直播业务的资源分配研究

提出一种车辆雾计算网络中视频直播业务的资源分配方法,通过联合优化比特率选择、用户调度和频谱资源分配,以实现在最大化视频质量的同时降低时延和视频抖动.为了降低时延和视频抖动,在效用函数的设计中将时延和比特率切换作为惩罚因子.由于网络的动态变化特性和可用的频谱资源,将上述优化问题建模为马尔可夫决策过程,采用Soft Actor-Critic深度强化学习算法获得最优资源分配策略.仿真结果证明,所提方法比现有强化学习算法具有更好的探索能力和收敛性能.     

基于认知无线电网络的协作中继资源分配算法

频谱资源紧张己经成为制约下一代无线通信系统发展的主要因素之一.认知无线电技术通过大力发掘授权频谱的复用潜力,可有效解决频谱紧缺问题.本文把协作中继传输技术引入到认知网络中,建立了配置双天线的次用户帮助主用户传输的协作模型,基于此提出了基于译码转发的协作功率及频谱资源分配算法,并进行了详细的理论分析.仿真结果表明,本文提出的协作传输策略能够在保障主用户正常传输情况下提高次用户的传输机会并减小中断概率.     

Deep Q-Learning Based Computation Offloading Strategy for Mobile Edge Computing

To reduce the transmission latency and mitigate the backhaul burden of the centralized cloud-based network services, the mobile edge computing (MEC) has been drawing increased attention from both industry and academia recently. This paper focuses on mobile users’ computation offloading problem in wireless cellular networks with mobile edge computing for the purpose of optimizing the computation offloading decision making policy. Since wireless network states and computing requests have stochastic properties and the environment’s dynamics are unknown, we use the model-free reinforcement learning (RL) framework to formulate and tackle the computation offloading problem. Each mobile user learns through interactions with the environment and the estimate of its performance in the form of value function, then it chooses the overhead-aware optimal computation offloading action (local computing or edge computing) based on its state. The state spaces are high-dimensional in our work and value function is unrealistic to estimate. Consequently, we use deep reinforcement learning algorithm, which combines RL method Q-learning with the deep neural network (DNN) to approximate the value functions for complicated control applications, and the optimal policy will be obtained when the value function reaches convergence. Simulation results showed that the effectiveness of the proposed method in comparison with baseline methods in terms of total overheads of all mobile users.     

A Comprehensive Utility Function for Resource Allocation in Mobile Edge Computing

In mobile edge computing (MEC), one of the important challenges is how much resources of which mobile edge server (MES) should be allocated to which user equipment (UE). The existing resource allocation schemes only consider CPU as the requested resource and assume utility for MESs as either a random variable or dependent on the requested CPU only. This paper presents a novel comprehensive utility function for resource allocation in MEC. The utility function considers the heterogeneous nature of applications that a UE offloads to MES. The proposed utility function considers all important parameters, including CPU, RAM, hard disk space, required time, and distance, to calculate a more realistic utility value for MESs. Moreover, we improve upon some general algorithms, used for resource allocation in MEC and cloud computing, by considering our proposed utility function. We name the improved versions of these resource allocation schemes as comprehensive resource allocation schemes. The UE requests are modeled to represent the amount of resources requested by the UE as well as the time for which the UE has requested these resources. The utility function depends upon the UE requests and the distance between UEs and MES, and serves as a realistic means of comparison between different types of UE requests. Choosing (or selecting) an optimal MES with the optimal amount of resources to be allocated to each UE request is a challenging task. We show that MES resource allocation is sub-optimal if CPU is the only resource considered. By taking into account the other resources, i.e., RAM, disk space, request time, and distance in the utility function, we demonstrate improvement in the resource allocation algorithms in terms of service rate, utility, and MES energy consumption.     

Resource Allocation in Edge-Computing Based Wireless Networks Based on Differential Game and Feedback Control

In this paper, we have proposed a differential game model to optimally solve the resource allocation problems in the edge-computing based wireless networks. In the proposed model, a wireless network with one cloud-computing center (CC) and lots of edge services providers (ESPs) is investigated. In order to provide users with higher services quality, the ESPs in the proposed wireless network should lease the computing resources from the CC and the CC can allocate its idle cloud computing resource to the ESPs. We will try to optimally allocate the edge computing resources between the ESPs and CC using the differential game and feedback control. Based on the proposed model, the ESPs can choose the amount of computing resources from the CC using feedback control, which is affected by the unit price of computing resources controlled by the CC. In the simulation part, the optimal allocated resources for users’ services are obtained based on the Nash equilibrium of the proposed differential game. The effectiveness and correctness of the proposed scheme is also verified through the numerical simulations and results.     

Task-Based Resource Allocation Bid in Edge Computing Micro Datacenter

Edge computing attracts online service providers (SP) to offload services to edge computing micro datacenters that are close to end users. Such offloads reduce packet-loss rates, delays and delay jitter when responding to service requests. Simultaneously, edge computing resource providers (RP) are concerned with maximizing incomes by allocating limited resources to SPs. Most works on this topic make a simplified assumption that each SP has a fixed demand; however, in reality, SPs themselves may have multiple task-offloading alternatives. Thus, their demands could be flexibly changed, which could support finer-grained allocations and further improve the incomes for RPs. Here, we propose a novel resource bidding mechanism for the RP in which each SP bids resources based on the demand of a single task (task-based) rather than the whole service (service-based) and then the RP allocates resources to these tasks with following the resource constraints at edge servers and the sequential rule of task-offloading to guarantee the interest of SPs. We set the incomes of the RP as our optimization target and then formulate the resource allocation problem. Two typical greedy algorithms are adopted to solve this problem and analyze the performance differences using two different bidding methods. Comprehensive results show that our proposal optimizes resource utilization and improves the RP’s incomes when resources in the edge computing datacenter are limited.     

Connected Vehicles Computation Task Offloading Based on Opportunism inCooperative Edge Computing

The traditional multi-access edge computing (MEC) capacity is overwhelmed by the increasing demand for vehicles, leading to acute degradation in task offloading performance. There is a tremendous number of resource-rich and idle mobile connected vehicles (CVs) in the traffic network, and vehicles are created as opportunistic ad-hoc edge clouds to alleviate the resource limitation of MEC by providing opportunistic computing services. On this basis, a novel scalable system framework is proposed in this paper for computation task offloading in opportunistic CV-assisted MEC. In this framework, opportunistic ad-hoc edge cloud and fixed edge cloud cooperate to form a novel hybrid cloud. Meanwhile, offloading decision and resource allocation of the user CVs must be ascertained. Furthermore, the joint offloading decision and resource allocation problem is described as a Mixed Integer Nonlinear Programming (MINLP) problem, which optimizes the task response latency of user CVs under various constraints. The original problem is decomposed into two subproblems. First, the Lagrange dual method is used to acquire the best resource allocation with the fixed offloading decision. Then, the satisfaction-driven method based on trial and error (TE) learning is adopted to optimize the offloading decision. Finally, a comprehensive series of experiments are conducted to demonstrate that our suggested scheme is more effective than other comparison schemes.     

Fuzzy Control Based Resource Scheduling in IoT Edge Computing

Edge Computing is a new technology in Internet of Things (IoT) paradigm that allows sensitive data to be sent to disperse devices quickly and without delay. Edge is identical to Fog, except its positioning in the end devices is much nearer to end-users, making it process and respond to clients in less time. Further, it aids sensor networks, real-time streaming apps, and the IoT, all of which require high-speed and dependable internet access. For such an IoT system, Resource Scheduling Process (RSP) seems to be one of the most important tasks. This paper presents a RSP for Edge Computing (EC). The resource characteristics are first standardized and normalized. Next, for task scheduling, a Fuzzy Control based Edge Resource Scheduling (FCERS) is suggested. The results demonstrate that this technique enhances resource scheduling efficiency in EC and Quality of Service (QoS). The experimental study revealed that the suggested FCERS method in this work converges quicker than the other methods. Our method reduces the total computing cost, execution time, and energy consumption on average compared to the baseline. The ES allocates higher processing resources to each user in case of limited availability of MDs; this results in improved task execution time and a reduced total task computation cost. Additionally, the proposed FCERS m 1m may more efficiently fetch user requests to suitable resource categories, increasing user requirements.     

Edge Computing-Based Tasks Offloading and Block Caching for Mobile Blockchain

Internet of Things (IoT) technology is rapidly evolving, but there is no trusted platform to protect user privacy, protect information between different IoT domains, and promote edge processing. Therefore, we integrate the blockchain technology into constructing trusted IoT platforms. However, the application of blockchain in IoT is hampered by the challenges posed by heavy computing processes. To solve the problem, we put forward a blockchain framework based on mobile edge computing, in which the blockchain mining tasks can be offloaded to nearby nodes or the edge computing service providers and the encrypted hashes of blocks can be cached in the edge computing service providers. Moreover, we model the process of offloading and caching to ensure that both edge nodes and edge computing service providers obtain the maximum profit based on game theory and auction theory. Finally, the proposed mechanism is compared with the centralized mode, mode A (all the miners offload their tasks to the edge computing service providers), and mode B (all the miners offload their tasks to a group of neighbor devices). Simulation results show that under our mechanism, mining networks obtain more profits and consume less time on average.     

User Profile System Based on Sentiment Analysis for Mobile Edge Computing

Emotions of users do not converge in a single application but are scattered across diverse applications. Mobile devices are the closest media for handling user data and these devices have the advantage of integrating private user information and emotions spread over different applications. In this paper, we first analyze user profile on a mobile device by describing the problem of the user sentiment profile system in terms of data granularity, media diversity, and server-side solution. Fine-grained data requires additional data and structural analysis in mobile devices. Media diversity requires standard parameters tointegrate user data from various applications. A server-side solution presents a potential risk when handling individual privacy information. Therefore, in order to overcome these problems, we propose a general-purposed user profile system based on sentiment analysis that extracts individual emotional preferences by comparing the difference between public and individual data based on particular features. The proposed system is built based on a sentiment hierarchy, which is created by using unstructured data on mobile devices. It can compensate for the concentration of single media, and analyze individual private data without the invasion of privacy on mobile devices.     

Machine Learning Based Resource Allocation of Cloud Computing in Auction

Resource allocation in auctions is a challenging problem for cloud computing. However, the resource allocation problem is NP-hard and cannot be solved in polynomial time. The existing studies mainly use approximate algorithms such as PTAS or heuristic algorithms to determine a feasible solution; however, these algorithms have the disadvantages of low computational efficiency or low allocate accuracy. In this paper, we use the classification of machine learning to model and analyze the multi-dimensional cloud resource allocation problem and propose two resource allocation prediction algorithms based on linear and logistic regressions. By learning a small-scale training set, the prediction model can guarantee that the social welfare, allocation accuracy, and resource utilization in the feasible solution are very close to those of the optimal allocation solution. The experimental results show that the proposed scheme has good effect on resource allocation in cloud computing.