Deep Reinforcement Learning for Delay-Oriented IoT Task Scheduling in SAGIN

Zhou, Conghao; Wu, Wen; He, Hua; Yang, Peng; Lyu, Feng; Cheng, Nan; Shen, Xuemin

doi:10.1109/twc.2020.3029143

Cited by 173 publications

(50 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yang et al dealt with resource allocation and task offloading problems to reduce the overall power consumption in their networks [7]. In another study, Zhou et al included satellite computation/communication as an alternative for UAVs [8]. They modelled their problem as a constrained Markov decision process (MDP) and provided a deep reinforcement learning solution.…”

Section: Related Workmentioning

confidence: 99%

“…4) ILP Solver: In addition to the heuristics we explained above, we use GUROBI Solver [11] to find the optimum solution for our multi-objective maximization problem (Eqs. [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. Despite the NP-hard property of our problem, that method could be used for only small solution space problems.…”

Section: A Q-learning Approachmentioning

confidence: 99%

See 1 more Smart Citation

Reinforcement Learning-Based Deadline and Battery-Aware Offloading in Smart Farm IoT-UAV Networks

Catherine¹,

Pamuklu²,

Syed³

et al. 2022

Preprint

View full text Add to dashboard Cite

Unmanned aerial vehicles (UAVs) with mounted base stations are a promising technology for monitoring smart farms. They can provide communication and computation services to extensive agricultural regions. With the assistance of a Multi-Access Edge Computing infrastructure, an aerial base station (ABS) network can provide an energy-efficient solution for smart farms that need to process deadline critical tasks fed by IoT devices deployed on the field. In this paper, we introduce a multiobjective maximization problem and a Q-Learning based method which aim to process these tasks before their deadline while considering the UAVs' hover time. We also present three heuristic baselines to evaluate the performance of our approaches. In addition, we introduce an integer linear programming (ILP) model to define the upper bound of our objective function. The results show that Q-Learning outperforms the baselines in terms of remaining energy levels and percentage of delay violations.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: A Q-learning Approachmentioning

confidence: 99%

Reinforcement Learning-Based Deadline and Battery-Aware Offloading in Smart Farm IoT-UAV Networks

Catherine¹,

Pamuklu²,

Syed³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…(1) DRL-based Computation Offloading in UAV-Assisted MEC Networks: Zhou et al [10] formulated the computing task scheduling problem as a constrained Markov decision process (CMDP) and solved it by proposing a novel risk-sensitive DRL method, where the UAV's energy consumption violation is defined as the risk metric. Liu et al [11] modified the vanilla Q-Learning algorithm to maximize the profit of the UAV under the constant cruising path.…”

Section: Related Work and Challengesmentioning

confidence: 99%

UAV-Assisted Privacy-Preserving Online Computation Offloading for Internet of Things

Wei

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Unmanned aerial vehicle (UAV) plays a more and more important role in Internet of Things (IoT) for remote sensing and device interconnecting. Due to the limitation of computing capacity and energy, the UAV cannot handle complex tasks. Recently, computation offloading provides a promising way for the UAV to handle complex tasks by deep reinforcement learning (DRL)-based methods. However, existing DRL-based computation offloading methods merely protect usage pattern privacy and location privacy. In this paper, we consider a new privacy issue in UAV-assisted IoT, namely computation offloading preference leakage, which lacks through study. To cope with this issue, we propose a novel privacy-preserving online computation offloading method for UAV-assisted IoT. Our method integrates the differential privacy mechanism into deep reinforcement learning (DRL), which can protect UAV’s offloading preference. We provide the formal analysis on security and utility loss of our method. Extensive real-world experiments are conducted. Results demonstrate that, compared with baseline methods, our method can learn cost-efficient computation offloading policy without preference leakage and a priori knowledge of the wireless channel model.

show abstract

“…LEO satellite network provides latitude of 160 km to 2000 km. To minimize end-to-end propagation delay, for better communication, wide coverage area and assured quality of services; many LEO satellite based commercial networks such as 'SpaceX', 'OneWeb', 'LeoSat' and 'Iridium' have come into existence [9,10]. LEO satellite based IoT services possess the potential of long-distance communication in remote areas.…”

Section: Introductionmentioning

confidence: 99%

Deep Reinforcement Learning-Based Long Short-Term Memory for Satellite IoT Channel Allocation

Durga¹,

Rajeshwari²,

Allehaibi³

et al. 2022

Intelligent Automation &Amp; Soft Computing

View full text Add to dashboard Cite

In recent years, the demand for smart wireless communication technology has increased tremendously, and it urges to extend internet services globally with high reliability, less cost and minimal delay. In this connection, low earth orbit (LEO) satellites have played prominent role by reducing the terrestrial infrastructure facilities and providing global coverage all over the earth with the help of satellite internet of things (SIoT). LEO satellites provide wide coverage area to dynamically accessing network with limited resources. Presently, most resource allocation schemes are designed only for geostationary earth orbit (GEO) satellites. For LEO satellites, resource allocation is challenging due to limited availability of resources. Moreover, due to uneven distribution of users on the ground, the satellite remains unaware of the users in each beam and therefore cannot adapt to changing state of users among the beams. In this paper, long short-term memory (LSTM) neural network has been implemented for efficient allocation of channels with the help of deep reinforcement learning (DRL) model. We name this model as DRL-LSTM scheme. Depending on the pool of resources available to the satellite, a channel allocation method based on the user density in each beam is designed. To make the satellite aware of the number of users in each beam, previous information related to the user density is provided to LSTM. It stores the information and allocates channels depending upon the requirement. Extensive simulations have been carried out which have shown that the DRL-LSTM scheme performs better as compared to the traditional and recently proposed schemes.

show abstract

Deep Reinforcement Learning for Delay-Oriented IoT Task Scheduling in SAGIN

Cited by 173 publications

References 37 publications

Reinforcement Learning-Based Deadline and Battery-Aware Offloading in Smart Farm IoT-UAV Networks

Reinforcement Learning-Based Deadline and Battery-Aware Offloading in Smart Farm IoT-UAV Networks

UAV-Assisted Privacy-Preserving Online Computation Offloading for Internet of Things

Deep Reinforcement Learning-Based Long Short-Term Memory for Satellite IoT Channel Allocation

Contact Info

Product

Resources

About