Cooperative Task Offloading and Block Mining in Blockchain-Based Edge Computing With Multi-Agent Deep Reinforcement Learning

Nguyen, Dinh C.; Ding, Ming; Pathirana, Pubudu N.; Seneviratne, Aruna; Li, Jun; Poor, H. Vincent

doi:10.1109/tmc.2021.3120050

Cited by 45 publications

(23 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After model consensus among ESs, an ES will be selected to work as a leader that builds an unverified block B and broadcasts it to all connected ESs and participating MDs for universal block mining. This selection can be based on the ESs' reputation [29] in the previous aggregation round via P2P collaboration, where the ES with the lowest data execution latency will be ranked with the best reputation to work as a leader in the current aggregation round. Here we focus on mining latency from the user side and thus the mining analysis at ESs is ignored.…”

Section: B Latency Of Block Mining In Bflmentioning

confidence: 99%

“…, where ξ is a pre-defined parameter of broadcasting verification result and comparison among network entities. Further, B|L M +N −1 | represents the average delay of the repeated verification on the block B of all entities except MD n [29]. Therefore, the mining latency at an MD can be given as…”

Section: B Latency Of Block Mining In Bflmentioning

confidence: 99%

“…We conduct numerical simulations to evaluate the performance of our proposed algorithm under various parameter settings. Inspired by related works [12], [14], [15], [24], [25], [29], we set up all necessary parameters for our BFL environment as listed in Table I.…”

Section: Simulations and Performance Evaluation A Parameter Settingsmentioning

confidence: 99%

“…We first investigate the training system reward (i.e., system utility as defined in (29)) performance by changing the learning rates at both actor and critic, which is important to determine a reliable parameter set for our later simulations, as shown in Fig. 7.…”

Section: Evaluation Of Drl Training Performancementioning

confidence: 99%

“…Thus, we will use learning rates γ a t = 0.003 and γ c t = 0.02 for the following simulations. Next, we compare the reward performance obtained by our proposed parameterized A2C algorithm against two baseline schemes: A2C with relaxed action space [29] and A2C with approximated action space [30]. For the first baseline scheme, we first relax the discrete offloading vector into a continuous set by defining a relaxed space as A a = {f (x 1 , ), f (x 2 ), ..., f (x N )}, where f (.)…”

Section: Evaluation Of Drl Training Performancementioning

confidence: 99%

See 4 more Smart Citations

Latency Optimization for Blockchain-Empowered Federated Learning in Multi-Server Edge Computing

Nguyen¹,

Hosseinalipour²,

Love³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this paper, we study a new latency optimization problem for Blockchain-based federated learning (BFL) in multiserver edge computing. In this system model, distributed mobile devices (MDs) communicate with a set of edge servers (ESs) to handle both machine learning (ML) model training and block mining simultaneously. To assist the ML model training for resource-constrained MDs, we develop an offloading strategy that enables MDs to transmit their data to one of the associated ESs. We then propose a new decentralized ML model aggregation solution at the edge layer based on a consensus mechanism to build a global ML model via peer-to-peer (P2P)-based Blockchain communications. We then formulate latency-aware BFL as an optimization aiming to minimize the system latency via joint consideration of the data offloading decisions, MDs' transmit power, channel bandwidth allocation for MDs' data offloading, MDs' computational allocation, and hash power allocation. To address the mixed action space of discrete offloading and continuous allocation variables, we propose a novel deep reinforcement learning scheme with a holistic design of a parameterized advantage actor critic (A2C) algorithm. Additionally, we theoretically characterize the convergence properties of the proposed BFL system in terms of the aggregation delay, mini-batch size, and number of P2P communication rounds. Our subsequent numerical evaluation demonstrates the superior performance of our proposed scheme over existing approaches in terms of model training efficiency, convergence rate, and system latency.

show abstract

Section: B Latency Of Block Mining In Bflmentioning

confidence: 99%

Section: B Latency Of Block Mining In Bflmentioning

confidence: 99%

Section: Simulations and Performance Evaluation A Parameter Settingsmentioning

confidence: 99%

Section: Evaluation Of Drl Training Performancementioning

confidence: 99%

Section: Evaluation Of Drl Training Performancementioning

confidence: 99%

See 3 more Smart Citations

Latency Optimization for Blockchain-Empowered Federated Learning in Multi-Server Edge Computing

Nguyen¹,

Hosseinalipour²,

Love³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

A blockchain‐enabled learning model based on distributed deep learning architecture

Zhang

Liang

Jia

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Aiming to address the unsatisfactory performance of existing distributed deep learning architectures, such as poor accuracy, slow network communication, low arithmetic speed, and insufficient security, we propose and design a learning model based on a distributed deep learning and blockchain architecture. We use a hybrid parallel algorithm based on blockchain (HP‐B) to build a distributed deep consensus learning model. The HP‐B algorithm is grouped according to the performance of computing nodes participating in training, network links and training samples, and the grouped computing equipment performs optimal distributed computing. The purpose of this approach is to solve the security and scalability concerns and improve the convergence speed and accuracy of deep learning. The proposed method achieves good results on the CIFAR‐100, CIFAR‐10, and IMAGENET data sets. Finally, the distributed deep learning model based on blockchain is combined with the generative adversarial network to solve the segmentation problem of medical imaging data, and the experimental results are superior to those of other networks.

show abstract

Secured SDN Based Task Scheduling in Edge Computing for Smart City Health Monitoring Operation Management System

Zhang,

Tang,

Wang

et al. 2023

J Grid Computing

View full text Add to dashboard Cite

Cooperative Task Offloading and Block Mining in Blockchain-Based Edge Computing With Multi-Agent Deep Reinforcement Learning

Cited by 45 publications

References 43 publications

Latency Optimization for Blockchain-Empowered Federated Learning in Multi-Server Edge Computing

Latency Optimization for Blockchain-Empowered Federated Learning in Multi-Server Edge Computing

A blockchain‐enabled learning model based on distributed deep learning architecture

Secured SDN Based Task Scheduling in Edge Computing for Smart City Health Monitoring Operation Management System

Contact Info

Product

Resources

About