Dynamic Computation Offloading and Resource Allocation for Multi-user Mobile Edge Computing

Nath, Samrat; Wu, Jingxian

doi:10.1109/globecom42002.2020.9348161

Cited by 21 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Peng et al [23] leveraged the deep deterministic policy gradient (DDPG) and hierarchical learning architectures to jointly solve the spectrum, computation and storage allocation problem in an EC based system. Recent studies [24], [25] solved the computation offloading and resource allocation problem for multiple mobile users in EC based systems by utilizing the DDPG-based framework and proposing the sate-of-the-art algorithms. Wang et al [26] jointly optimized the communication, computation and caching resources problem in an EC scenario by utilizing a twin-actor DDPG framework.…”

Section: B Machine Learning-based Methodsmentioning

confidence: 99%

“…Suppose C represents the number of CPU cycles required to process 1 Byte of data, then L u is given as the total CPU cycles required to compute data d u (L u = d u × C). A similar task computation model (with CPU cycles) is proposed in [24], [25] as well.…”

Section: B Users and Jobs Modelmentioning

confidence: 99%

“…The ability of Machine Learning (ML) based approaches to adapt to the changes in non-stationary environments has attracted increasingly more research interests. Among ML-based techniques, deep-learning (DL) based methods [21], [22], [23], [24], [25], [26] have gained significant popularity to solve the control and management problems in IoT ecosystem. Deep deterministic policy gradient (DDPG) based methods are adopted when coping with dynamic optimization problems in a continuous action space [23], [24], [25], [26].…”

Section: Introductionmentioning

confidence: 99%

“…Among ML-based techniques, deep-learning (DL) based methods [21], [22], [23], [24], [25], [26] have gained significant popularity to solve the control and management problems in IoT ecosystem. Deep deterministic policy gradient (DDPG) based methods are adopted when coping with dynamic optimization problems in a continuous action space [23], [24], [25], [26]. More details can be found in Section II.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Deep-Deterministic Policy Gradient Based Multi-Resource Allocation in Edge-Cloud System: A Distributed Approach

Qadeer

Lee

2023

IEEE Access

View full text Add to dashboard Cite

Edge Cloud (EC) empowers the beyond 5G (B5G) wireless networks to cope with large-scale and real-time traffics of Internet-of-Things (IoT) by minimizing the latency and providing compute power at the edge of the network. Due to a limited amount of resources at the EC compared to the backend cloud (BC), intelligent resource management techniques become imperative. This paper studies the problem of multi-resource allocation (MRA) in terms of compute and wireless resources in an integrated EC and BC environment. Machine learning-based approaches are emerging to solve such optimization problems. However, it is challenging to adopt traditional discrete action space-based methods due to their high dimensionality issue. To this end, we propose a deep-deterministic policy gradient (DDPG) based temporal feature learning attentional network (TFLAN) model to address the MRA problem. TFLAN combines convolution, gated recurrent unit and attention layers together to mine local and long term temporal information from the task sequences for excellent function approximation. A novel heuristic-based priority experience replay (hPER) method is formulated to accelerate the convergence speed. Further, a pruning principle helps the TFLAN agent to significantly reduce the computational complexity and balance the load among base stations and servers to minimize the rejection-rate. Lastly, data parallelism technique is adopted for distributed training to meet the needs of a high-volume of IoT traffic in the EC environment. Experimental results demonstrate that the distributed training approach suites well to the problem scale and can magnify the speed of the learning process. We validate the proposed framework by comparing with five state-of-theart RL agents. Our proposed agent converges fast and achieves up to 28% and 72% reduction in operational cost and rejection-rate, and achieves up to 32% gain in the quality of experience on average, compared to the most advanced DDPG agent.

show abstract

Section: B Machine Learning-based Methodsmentioning

confidence: 99%

Section: B Users and Jobs Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Deep-Deterministic Policy Gradient Based Multi-Resource Allocation in Edge-Cloud System: A Distributed Approach

Qadeer

Lee

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Computation offloading strategies have been studied extensively [116][117][118][119] and generally fall under two categories, full offloading (coarse-grained) [120] and partial offloading (fine-grained) [91], [121][122][123][124]. For coarse-grained mode, the data set of a task has to be executed as a whole either locally or remotely on an edge server.…”

Section: Offloading Strategiesmentioning

confidence: 99%

A Review of the Current Task Offloading Algorithms, Strategies and Approach in Edge Computing Systems

Acheampong¹,

Zhang²,

Xu³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

Task offloading is an important concept for edge computing and the Internet of Things (IoT) because computationintensive tasks must be offloaded to more resource-powerful remote devices. Task offloading has several advantages, including increased battery life, lower latency, and better application performance. A task offloading method determines whether sections of the full application should be run locally or offloaded for execution remotely. The offloading choice problem is influenced by several factors, including application properties, network conditions, hardware features, and mobility, influencing the offloading system's operational environment. This study provides a thorough examination of current task offloading and resource allocation in edge computing, covering offloading strategies, algorithms, and factors that influence offloading. Full offloading and partial offloading strategies are the two types of offloading strategies. The algorithms for task offloading and resource allocation are then categorized into two parts: machine learning algorithms and non-machine learning algorithms. We examine and elaborate on algorithms like Supervised Learning, Unsupervised Learning, and Reinforcement Learning (RL) under machine learning. Under the non-machine learning algorithm, we elaborate on algorithms like non(convex) optimization, Lyapunov optimization, Game theory, Heuristic Algorithm, Dynamic Voltage Scaling, Gibbs Sampling, and Generalized Benders Decomposition (GBD). Finally, we highlight and discuss some research challenges and issues in edge computing.

show abstract

Computation offloading techniques in edge computing: A systematic review based on energy, QoS and authentication

Kanupriya,

Chana,

Goyal

2024

Concurrency and Computation

View full text Add to dashboard Cite

SummaryIn today's era, Internet of Things (IoT) devices generate a vast amount of data, which is typically stored in the cloud environment and can be accessed by edge and IoT devices. The data generated by these devices are offloaded through computation offloading (CO) techniques in an edge/cloud computing environment. This paper conducts a systematic literature review (SLR) to review the state‐of‐the‐art CO techniques in edge computing (EC) in the context of energy, Quality of Service (QoS), authentication and traceability. In this SLR, the evolution of offloading techniques is analyzed in detail. A total of 138 articles, spanning from 2016 to 2023 (till date), have been classified into QoS, energy, and authentication and traceability‐based CO techniques. The optimization‐based techniques are the most preferred choices to improve the QoS and reduce energy in the research field of CO in EC. In addition, this paper explores the significant issues and challenges that require further investigation. For future research, energy, QoS, and data provenance for dependent or dynamic task offloading in mobility scenarios can be explored further.

show abstract

Dynamic Computation Offloading and Resource Allocation for Multi-user Mobile Edge Computing

Cited by 21 publications

References 14 publications

Deep-Deterministic Policy Gradient Based Multi-Resource Allocation in Edge-Cloud System: A Distributed Approach

Deep-Deterministic Policy Gradient Based Multi-Resource Allocation in Edge-Cloud System: A Distributed Approach

A Review of the Current Task Offloading Algorithms, Strategies and Approach in Edge Computing Systems

Computation offloading techniques in edge computing: A systematic review based on energy, QoS and authentication

Contact Info

Product

Resources

About