Learning Driven Computation Offloading for Asymmetrically Informed Edge Computing

Hu, Miao; Zhuang, Lei; Wu, Di; Zhou, Yipeng; Chen, Xu; Xiao, Liang

doi:10.1109/tpds.2019.2893925

Cited by 64 publications

(13 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To evaluate the performance of X-DDRL in a real-world scenario, we created a testbed, similar to [31], [43]. The type of servers are the same as simulation setup while the number of servers of each type is as follows: two Raspberry pi 3B, one Raspberry pi 4B, one Jetson Nano, one instance of Horizon Cloud, and six m3.large instances of Nectar Cloud infrastructure).…”

Section: Testbed Setupmentioning

confidence: 99%

A Distributed Deep Reinforcement Learning Technique for Application Placement in Edge and Fog Computing Environments

Goudarzi¹,

Palaniswami²,

Buyya³

2021

Preprint

View full text Add to dashboard Cite

Fog/Edge computing is a novel computing paradigm supporting resource-constrained Internet of Things (IoT) devices by placement of their tasks on edge and/or cloud servers. Recently, several Deep Reinforcement Learning (DRL)-based placement techniques have been proposed in fog/edge computing environments, which are only suitable for centralized setups. The training of well-performed DRL agents requires manifold training data while obtaining training data is costly. Hence, these centralized DRL-based techniques lack generalizability and quick adaptability, thus failing to efficiently tackle application placement problems. Moreover, many IoT applications are modeled as Directed Acyclic Graphs (DAGs) with diverse topologies. Satisfying dependencies of DAG-based IoT applications incur additional constraints and increase the complexity of placement problem. To overcome these challenges, we propose an actor-critic-based distributed application placement technique, working based on the IMPortance weighted Actor-Learner Architectures (IMPALA). IMPALA is known for efficient distributed experience trajectory generation that significantly reduces exploration costs of agents. Besides, it uses an adaptive off-policy correction method for faster convergence to optimal solutions. Our technique uses recurrent layers to capture temporal behaviors of input data and a replay buffer to improve the sample efficiency. The performance results, obtained from simulation and testbed experiments, demonstrate that our technique significantly improves execution cost of IoT applications up to 30% compared to its counterparts.

show abstract

Section: Testbed Setupmentioning

confidence: 99%

A Distributed Deep Reinforcement Learning Technique for Application Placement in Edge and Fog Computing Environments

Goudarzi¹,

Palaniswami²,

Buyya³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…From this point of view, the factors that affect the offloading performance include wireless channel conditions, wireless bandwidth, and processing capability of the destination node (i.e., the node to which the task is offloaded). The research on binary offloading involves in the association between tasks and destination nodes [102]- [106], which refers to the determination of the offloading of a specific task to a destination node, among various tasks and destination nodes.…”

Section: B2 Granularitymentioning

confidence: 99%

Resource Scheduling in Edge Computing: A Survey

Luo

et al. 2021

Preprint

View full text Add to dashboard Cite

With the proliferation of the Internet of Things (IoT) and the wide penetration of wireless networks, the surging demand for data communications and computing calls for the emerging edge computing paradigm. By moving the services and functions located in the cloud to the proximity of users, edge computing can provide powerful communication, storage, networking, and communication capacity. The resource scheduling in edge computing, which is the key to the success of edge computing systems, has attracted increasing research interests. In this paper, we survey the state-of-the-art research findings to know the research progress in this field. Specifically, we present the architecture of edge computing, under which different collaborative manners for resource scheduling are discussed. Particularly, we introduce a unified model before summarizing the current works on resource scheduling from three research issues, including computation offloading, resource allocation, and resource provisioning. Based on two modes of operation, i.e., centralized and distributed modes, different techniques for resource scheduling are discussed and compared. Also, we summarize the main performance indicators based on the surveyed literature. To shed light on the significance of resource scheduling in real-world scenarios, we discuss several typical application scenarios involved in the research of resource scheduling in edge computing. Finally, we highlight some open research challenges yet to be addressed and outline several open issues as the future research direction.

show abstract

“…However, traditional computation model does not work well in edge computing since the three main reasons: i) UEs are heterogeneous in practical edge computing scenarios, ii) the configures of equipment used by each edge user are unseen to server-side, and iii) the algorithms (e.g. CNN/DQN/RNN) used for tasks also impact the diversity of execution efficiency [26]. Therefore, we propose a time-based computation model to achieve better abstraction in soCoM system.…”

Section: A Computation Modelmentioning

confidence: 99%

Semi-Online Computational Offloading by Dueling Deep-Q Network for User Behavior Prediction

et al. 2020

View full text Add to dashboard Cite

Task offloading could optimize computational resource utilization in edge computing environments. However, how to assign and offload tasks for different behavior users is an essential problem since the systems dynamic, intelligent application diversity, and user personality. With user behavior prediction, this paper proposes soCoM, a semi-online Computational Offloading Model. We explore the user behaviors in sophisticated action space by reinforcement learning for catching unknown environment information. With Dueling Deep-Q Network, both the prediction accuracy of users' behaviors and the server load balance are well-considered, while increasing the computational efficiency and decreasing the resource costing. We propose a dynamic simulation environment of edge computing to demonstrate that user behavior is the critical factor for impacting system utilization. As the action space increasing, Dueling DQN performs better than state-of-art DQN and other improved strategies, and also load balance in multiple different server scenario.

show abstract

Learning Driven Computation Offloading for Asymmetrically Informed Edge Computing

Cited by 64 publications

References 39 publications

A Distributed Deep Reinforcement Learning Technique for Application Placement in Edge and Fog Computing Environments

A Distributed Deep Reinforcement Learning Technique for Application Placement in Edge and Fog Computing Environments

Resource Scheduling in Edge Computing: A Survey

Semi-Online Computational Offloading by Dueling Deep-Q Network for User Behavior Prediction

Contact Info

Product

Resources

About