Deep Reinforcement Learning-Based Workload Scheduling for Edge Computing

Zheng, Tao; Wan, Jian; Zhang, Jilin; Jiang, Congfeng

doi:10.1186/s13677-021-00276-0

Cited by 47 publications

(15 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Equation (20), the propensity of the solution to approach the best solution is shown by the second term, while the tendency to avoid the worst solution is indicated by the third term. The placement of individuals in the population is updated for each iteration using two terms in the proposed hybrid technique.…”

Section: Hybrid Jaya-gray Wolf Optimization (Hyjgwo) Algorithm: the P...mentioning

confidence: 99%

“…More specifically, the update will be carried out by Equations ( 17), (18), and (19) as in the Gray wolf optimization, if the wolf is close to alpha, beta, or delta. Otherwise, the proposed hybrid Jaya-gray wolf optimizer (HyJGWO) evaluates the position update equation as given in Equation (20). This change will aid in overcoming the issues of delayed convergence and trapping into local optima.…”

Section: Hybrid Jaya-gray Wolf Optimization (Hyjgwo) Algorithm: the P...mentioning

confidence: 99%

See 1 more Smart Citation

An ML‐based task clustering and placement using hybrid Jaya‐gray wolf optimization in fog‐cloud ecosystem

Keshri,

Vidyarthi

2024

Concurrency and Computation

View full text Add to dashboard Cite

SummaryThe rapid expansion of IoT systems has caused network congestion and delays in task placement and resource provisioning as usually the tasks are executed at a far location in the cloud. Fog computing reduces the computing burden of cloud data centers as well as the communication burden of the internet as fog resources are placed near the data generation points. Within Fog computing, an important challenge is the optimal task placement which is an NP‐class problem. This work applies machine learning for task clustering and addresses the task placement problem in a fog computing environment using a hybrid of two recent metaheuristics; Jaya and gray wolf optimization (GWO). The hybrid method considers optimizing the total number of active fog nodes, load balancing in fog nodes, and average response time of the tasks. The performance of the proposed method is evaluated on a real‐time LCG dataset and is compared with reinforcement learning fog scheduling (RLFS), genetic algorithm (GA), dynamic resource allocation mechanism (DRAM), load balancing and scheduling algorithm (LBSSA), and particle swarm optimization with simulated annealing (PSO‐SA) algorithms. The results demonstrate the superiority of the suggested method over the baseline techniques in terms of average improvement of 51.04% in load balance variance, 30.25% in average response time, 24.16% in execution time, and 47.10% in the number of devices used.

show abstract

Section: Hybrid Jaya-gray Wolf Optimization (Hyjgwo) Algorithm: the P...mentioning

confidence: 99%

Section: Hybrid Jaya-gray Wolf Optimization (Hyjgwo) Algorithm: the P...mentioning

confidence: 99%

An ML‐based task clustering and placement using hybrid Jaya‐gray wolf optimization in fog‐cloud ecosystem

Keshri,

Vidyarthi

2024

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…We may accomplish both goals by aligning the use of cloud resources with the previous one. To intuitively deliver cloud resources ahead of demand, this study utilizes an automated dynamic resource allocation system that is based on machine learning algorithms (Zhang et al, 2022). When analyzing heuristic data from resource utilization when users use certain applications, this technique considers user settings.…”

Section: Bodymentioning

confidence: 99%

“…To determine whether the host was overloaded, the majority of the analysis was on its current CPU time utilization. If changes to host energy modes and relocations of virtual machines are not essential, consolidation attempts could be slowed down (Zhang et al, 2022).…”

Section: Allocation Of Resourcesmentioning

confidence: 99%

Optimizing Resource Allocation in Cloud Computing Using Machine Learning

Jeyaraman,

Bayani,

Malaiyappan

2024

EJT

View full text Add to dashboard Cite

Purpose: A key component of large-scale distributed computing is the allocation of resources, as computer networks cooperate to address complex optimization problems. To get the most out of computers in general, or throughput, is the goal of resource allocation in this case. When it comes to distributed computing, there are two main varieties: grid computing and cloud computing. In grid computing, many geographically dispersed clusters are linked and made available to the general public. Materials and Methods: We looked at Monte Carlo Tree Search and Long-Short Term Memory and examined how efficient they were. By maintaining consistent traffic patterns, the simulation demonstrated that MCTS performance was improved. However, such a strategy is difficult to implement due to the potential speed with which traffic patterns may alter. A successful service level agreement (SLA) was achieved, and the issue was shown to be fixable using LSTM. We compare the suggested model to different load-balancing algorithms to find the one that best allocates resources. Findings: The results show that compared to the state-of-the-art models, the suggested model achieves an accuracy rate that is 10-15% higher. The suggested model lowers the error percentage rate of the average request blocking likelihood of traffic load by around 9.5-10.2% when compared to the predictions of existing models. Therefore, the proposed method has the potential to enhance network utilization by reducing the amount of time required by memory and the central processing unit. Implications to Theory, Practice and Policy: One advantage of the new method is a more robust forecasting strategy in comparison to earlier models. Using firefly algorithms, future research will construct a cloud data center that employs a variety of heuristics and machine learning methodologies to load balance the energy cloud (Oshawa et al., 2022).

show abstract

“…During task allocation, the execution strategy corresponding to the solution x needs to be solved to meet the dependencies between tasks and environmental parameters, that is, the solution obtained should be in the feasible space. In the research of task allocation, the existing deep reinforcement learning methods usually regard it as an end-to-end learning task, and have designed different models and training methods [11][12][13][14][15][16][17]. However, by exploring each step of action, the model will not get a reward function value for completing the task until the whole scheduling task is completed, resulting in sparse rewards, large state space, and difficulty in training.…”

Section: Graph Convolution Fusion Scheduling Modelmentioning

confidence: 99%

Task Scheduling Strategy of Logistics Cloud Robot Based on edge computing

Tang,

Jiao,

Xue

et al. 2023

Preprint

View full text Add to dashboard Cite

The logistics industry has entered the era of highly information-based intelligent logistics. In order to save resource consumption and improve the cooperation efficiency between robots, a centralized computing mode has emerged in the application of warehousing robots, namely, logistics cloud robots. In order to improve the work efficiency of the logistics cloud robot, this paper designs a task resource scheduling model of the cloud logistics robot in the edge computing environment. The logistics cloud robot can improve the response efficiency by unloading some delay sensitive tasks to the edge server, and assign the delay insensitive tasks to the cloud computing center. Therefore, it is particularly important to reasonably allocate tasks to appropriate computing cells. The model combines graph convolution network and heuristic algorithm, and designs a joint scheduling strategy to unload requests to edge services and cloud servers reasonably, so as to maximize the use of computing resources. Based on the joint scheduling strategy, the optimal long-term scheduling decision is generated to reduce scheduling time and improve service quality. To evaluate the effectiveness of the model, we use heterogeneous earliest completion time (HEFT) and other algorithms as benchmarks to compare the performance of the model. In order to verify the authenticity of the scheduling scheme, we conducted simulation experiments on the EdgeCloudSim platform to verify the quality of service of task scheduling in different scenarios. The experimental results show that the proposed method is superior to most of the comparison methods and can better optimize the task scheduling efficiency of the logistics cloud robot.

show abstract

Deep Reinforcement Learning-Based Workload Scheduling for Edge Computing

Cited by 47 publications

References 17 publications

An ML‐based task clustering and placement using hybrid Jaya‐gray wolf optimization in fog‐cloud ecosystem

An ML‐based task clustering and placement using hybrid Jaya‐gray wolf optimization in fog‐cloud ecosystem

Optimizing Resource Allocation in Cloud Computing Using Machine Learning

Task Scheduling Strategy of Logistics Cloud Robot Based on edge computing

Contact Info

Product

Resources

About