关键词： Mobile edge computing   Dynamic service migration   Reinforcement learning   Q-learning   Software-defined networking  

摘要： Service migration in mobile edge computing is a promising approach to improving the quality of service (QoS) for mobile users and reducing the network operational cost for service providers as well. However, these benefits are not free, coming at costs of bulk-data transfer, and likely service disruption, which could consequently increase the overall service costs. To gain the benefits of service migration while minimizing its cost across the edge nodes, in this paper, we leverage reinforcement learning (RL) method to design a cost-effective framework, called Mig-RL, for the service migration with a reduction of total service costs as a goal in a mobile edge environment. The Mig-RL leverages the infrastructure of edge network and deploys a migration agent through Q-leaming to learn the optimal policy with respect to the service migration status. We distinguish the Mig-RL from other existing works in several major aspects. First, we fully exploit the nature of this problem in a modest migration space, which allows us to constrain the number of service replicas whereby a defined state-action space could be effectively handled, as opposed to those methods that need to always approximate a huge state-action space for policy optimality. Second, we advocate a migration policy-base as a cache to save the learning process by retrieving the most effective policy whenever a similar migration pattern is encountered as time goes on. Finally, by exploiting the idea of software defined network, we also investigate the efficient implementation of Mig-RL in mobile edge network. Experimental results based on some real and synthesized access sequences show that Mig-RL, compared with the selected existing algorithms, can substantially minimize the service costs, and in the meantime, efficiently improve the QoS by adapting to the changes of mobile access patterns. (C) 2020 Elsevier Inc. All rights reserved.

关键词： cloud computing   mobile computing   stochastic processes   wireless channels   minimisation   resource allocation   power aware computing   network servers   telecommunication power management   computer network performance evaluation   multitier MEC offloading strategy   dynamic channel characteristics   cloud computing   mobile edge computing   edge servers   computation task   computing capacity   three‐   layer network architecture   mobile devices   edge cloudlets   helper cloudlet   computing data   joint communication   computation resource allocation problem   energy consumption   closed‐   form solutions   optimal offloading data size   optimisation problem   cloud execution   data transmission rate   stochastic channel condition   mobile execution   dynamic voltage and frequency scaling technique   network stability improvement   computation cost minimisation  

摘要： Comparing with the cloud computing, mobile edge computing (MEC) can further decrease the latency and improve the stability of the networks. However, it is challenging for the edge servers to deal with the large computation task due to the limited computing capacity. In this study, we design a novel three-layer network architecture consisting of mobile devices, edge cloudlets, and helper cloudlets, where the computing data can be partially processed at the edge cloudlet and helper cloudlet. Based on this, a joint communication, offloading, and computation resource allocation problem is formulated to minimise the computation cost and energy consumption. Due to its difficulty to directly solve the formulated problem, we first propose an offloading scheme to obtain the closed-form solutions for the optimal offloading data size. Next, we decompose the optimisation problem into two subproblems: (i) for the cloud execution, we dynamically adjust the data transmission rate according to the stochastic channel condition, (ii) for the mobile execution, the energy consumption can be further reduced by applying the dynamic voltage and frequency scaling technique. Finally, the numerical results demonstrate the efficiency of the proposed scheme, and show the performance gains in terms of delay, computation cost and energy consumption.

关键词： mobile edge computing   wireless power transmission   convex optimization   augmented Lagrangian method  

摘要： With the rapid popularity of mobile devices (MDs), mobile edge computing (MEC) networks and wireless power transmission (WPT) will receive more attention. Naturally, by integrating these two technologies, the inherent energy consumption during task execution can be effectively reduced, and the collected energy can be provided to charge the MD. In this article, our research focuses on extending the battery time of MDs by maximizing the harvested energy and minimizing the consumed energy in the MEC-WPT system, which is formulated as a residual energy maximization problem and also a non-convex optimization problem. On the basis of study on maximizing the residual energy under multi-users and multi-time blocks, we propose an effective jointly optimization method (i.e., jointly optimize the energy harvesting time, task-offloading time, task-offloading size and the MDs' CPU frequency), which combines the convex optimization method and the augmented Lagrangian to solve the residual energy maximum problem. We leverage Time Division Multiple Access (TMDA) mode to coordinate computation offloading. Simulation results show that our scheme has better performance than the benchmark schemes on maximizing residual energy. In particular, our proposed scheme is outstanding in the failure rate of multiple MDs and can adapt to the task size to minimize the failure rate.

关键词： Mobile edge computing   colocation   bandwidth trading   auction  

摘要： Inspired by the shared infrastructure of colocation data centers and the growth of Mobile Edge Computing (MEC), colocation MEC businesses have thrived to offer an economical and low latency solution for MEC based tenants. In colocation MEC, a colocation service provider leases spaces at the base stations (BSs) to tenants for housing their servers. Although the tenants fully manage their own servers, they still need to purchase bandwidth from the network provider to via bandwidth purchasing market to serve their users. As a result, tenants should consider the trade-off between delay and energy consumption, which are related to bandwidth and server computing speed, respectively. In this paper, an auction framework is used to model the interaction between the network provider and tenants in the bandwidth market. The network provider determines the bandwidth amount and corresponding payment by tenants based on their submitted bids, in which the valuation of the bandwidth amount in each bid is based on the coupling between bandwidth and server computing speed. The networking provider also seeks winners in the social welfare maximization problem, which is NP-hard. To solve this problem, we propose a solution based on the randomized auction mechanism, which is proved to be approximately truthful, individually rational, and computationally efficient. Simulation results verify the effectiveness of our proposed randomized auction scheme in bandwidth trading.

关键词： Task analysis   Trajectory   Unmanned aerial vehicles   Resource management   Servers   Computer architecture   Relays   Mobile edge computing   UAV   wireless power transfer   resource allocation   trajectory design  

摘要： A wireless-powered mobile edge computing (MEC) architecture with the cooperation between an access point (AP) and an unmanned aerial vehicle (UAV) is studied in this article. The AP, powered by the grid, is integrated with a high-performance processing server to help compute the user equipment's (UEs') offloaded tasks while also performing high-power laser-like energy charging for the UAV. The UAV serves as (1) an information relay to help the UEs offload/download their computation tasks/results, (2) an energy relay to broadcast energy from the AP to the UEs, as well as (3) an MEC server to help the UEs compute their tasks. We aim at maximizing the weighted sum completed task-input bits (WSCTB) of UEs under the task and time allocation, information-causality, energy-causality, and the UAV's trajectory constraints, by jointly optimizing the task and time allocation as well as the UAV's energy transmit power and trajectory. The formulated WSCTB maximization problem is non-convex, and we propose a three-step block coordinate descending algorithm to address three sub-problems iteratively for obtaining a proper solution. Simulation results show that the UAV's trajectories highly depend on the AP's location and the UEs' weight values. In addition, significant performance improvement is achieved by the proposed algorithm compared to some practical benchmarks.

关键词： Task analysis   Data mining   Data processing   Servers   Privacy   Mobile handsets   Blockchain   mobile edge computing   mining   task offloading   privacy   deep reinforcement learning  

摘要： Blockchain technology with its secure, transparent and decentralized nature has been recently employed in many mobile applications. However, the process of executing extensive tasks such as computation-intensive data applications and blockchain mining requires high computational and storage capability of mobile devices, which would hinder blockchain applications in mobile systems. To meet this challenge, we propose a mobile edge computing (MEC) based blockchain network where multi-mobile users (MUs) act as miners to offload their data processing tasks and mining tasks to a nearby MEC server via wireless channels. Specially, we formulate task offloading, user privacy preservation and mining profit as a joint optimization problem which is modelled as a Markov decision process, where our objective is to minimize the long-term system offloading utility and maximize the privacy levels for all blockchain users. We first propose a reinforcement learning (RL)-based offloading scheme which enables MUs to make optimal offloading decisions based on blockchain transaction states, wireless channel qualities between MUs and MEC server and user's power hash states. To further improve the offloading performances for larger-scale blockchain scenarios, we then develop a deep RL algorithm by using deep Q-network which can efficiently solve large state space without any prior knowledge of the system dynamics. Experiment and simulation results show that the proposed RL-based offloading schemes significantly enhance user privacy, and reduce the energy consumption as well as computation latency with minimum offloading costs in comparison with the benchmark offloading schemes.

关键词： Mobile edge computing   computation offloading   demand response   energy storage control   renewable usage  

摘要： Mobile Edge Computing (MEC) networks have recently developed to integrate renewable sources for providing sustainable and low-cost computation/energy services. However, one challenging problem is how to guarantee the stable and effective services for numerous devices under highly intermittent renewable generations and variable computation/energy demands. In this paper, we consider to co-deploy the edge servers and energy storage at micro Base Station (BS) in each edge sub-network. In the considered scenario, we first determine the number of edge servers and energy storage units at different sub-edge networks in a long-term period. Then, we jointly design the offloading and demand response policies for all micro BSs in a centralized manner. Considering the variabilities of renewable energy and computation/energy demands, the system cost is minimized while meeting the energy demand-supply balance and renewable portfolio standards. Furthermore, a distributed optimization problem is formulated to find the optimal offloading parameters, the amount of charging/discharging energy and utilized renewable energy of the micro BS without handing over all operational information. Numerical results show that the proposed deployment and policies can provide effective offloading and energy saving strategies for the proposed mobile edge network.

关键词： Deep reinforcement learning   Intelligent IoT   Mobile edge computing  

摘要： In this paper, we investigate mobile edge computing (MEC) networks for intelligent internet of things (IoT), where multiple users have some computational tasks assisted by multiple computational access points (CAPs). By offloading some tasks to the CAPs, the system performance can be improved through reducing the latency and energy consumption, which are the two important metrics of interest in the MEC networks. We devise the system by proposing the offloading strategy intelligently through the deep reinforcement learning algorithm. In this algorithm, Deep Q-Network is used to automatically learn the offloading decision in order to optimize the system performance, and a neural network (NN) is trained to predict the offloading action, where the training data is generated from the environmental system. Moreover, we employ the bandwidth allocation in order to optimize the wireless spectrum for the links between the users and CAPs, where several bandwidth allocation schemes are proposed. In further, we use the CAP selection in order to choose one best CAP to assist the computational tasks from the users. Simulation results are finally presented to show the effectiveness of the proposed reinforcement learning offloading strategy. In particular, the system cost of latency and energy consumption can be reduced significantly by the proposed deep reinforcement learning based algorithm. (c) 2020 Elsevier B.V. All rights reserved.

关键词： Federated learning   Mobile edge computing   Deep neural network   Model partition   Differential privacy  

摘要： Federated learning is a recently proposed paradigm that presents significant advantages in privacy-preserving machine learning services. It enables the deep learning applications on mobile devices, where a deep neural network (DNN) is trained in a decentralized manner among thousands of edge clients. However, directly apply the federated learning algorithm to the mobile edge computing environment will incur unacceptable computation costs in mobile edge devices. Moreover, among the training process, frequent model parameters exchanging between participants and the central server will increase the leakage possibility of the users' sensitive training data. Aiming at reducing the heavy computation cost of DNN training on edge devices while providing strong privacy guarantees, we propose a mobile edge computing enabled federated learning framework, called FedMEC, which integrating model partition technique and differential privacy simultaneously. In FedMEC, the most complex computations can be outsourced to the edge servers by splitting a DNN model into two parts. Furthermore, we apply the differentially private data perturbation method to prevent the privacy leakage from the local model parameters, in which the updates from an edge device to the edge server is perturbed by the Laplace noise. To validate the proposed FedMEC, we conduct a series of experiments on an image classification task under the settings of federated learning. The results demonstrate the effectiveness and practicality of our FedMEC scheme.

关键词： Task analysis   Edge computing   Processor scheduling   Servers   Wireless communication   Delays   Dynamic scheduling   Mobile edge computing   task offloading   Internet-of-Things   LTE over unlicensed   constrained deep reinforcement learning  

摘要： To facilitate the deployment of private industrial Internet-of-Things (IoT), applying long-term-evolution (LTE) over unlicensed spectrum (LTE-U) is a promising technology, which can deal with the licensed spectrum scarcity problem and the stringent quality-of-service (QoS) requirement via centralized control. In this paper, we investigate the computing offloading problem for LTE-U-enabled IoT, where computing tasks on an IoT device are either executed locally or offloaded to the edge server on an LTE-U base station. Considering a constrained edge computing cost (e.g., operation power consumption) for offloaded tasks, the task scheduling problem is formulated as a constrained Markov decision process (CMDP) to maximize the long-term average reward, which integrates both task completion profit and task completion delay. In order to address the uncertainty of task arrivals and channel availability, a constrained deep Q-learning-based task scheduling algorithm with provable convergence is proposed, where an adaptive reward function can appropriately bound the average edge computing cost. Extensive simulation results show that the proposed scheme considerably enhances the system performance.