Deep reinforcement learning for computation offloading in mobile edge computing environment

Chen, Miaojiang; Wang, Tian; Zhang, Shaobo; Liu, Anfeng

doi:10.1016/j.comcom.2021.04.028

Cited by 62 publications

(18 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mobile-edge computing (MEC) offers execution resources such as storage, computations, etc., close to the users (in a network), that can be utilized to deliver services, as well as store and process the content. Artificial intelligence techniques help to further improve the performance of MEC [ 85 , 86 ]. Fifth-generation and MEC technologies together have the potential to greatly enhance performance and allow the real-time processing of large volumes of data.…”

Section: Insights Into Communication Standards and Technologiesmentioning

confidence: 99%

Communication Technologies for Interoperable Smart Microgrids in Urban Energy Community: A Broad Review of the State of the Art, Challenges, and Research Perspectives

Reddy

Kumar

Chakravarthi

2022

Sensors

View full text Add to dashboard Cite

In modern urban energy communities, diverse natured loads (homes, schools, hospitals, malls, etc.) are situated in the same locality and have self-electricity generation/management facilities. The power systems of these individual buildings are called smart microgrids. Usually, their self-electricity generation is based on renewable energy sources, which are uncertain due to their environmental dependency. So, the consistency of self-energy generation throughout the day is not guaranteed; thus, the dependency on the central utility grid is continued. To solve this, researchers have recently started working on interoperable smart microgrids (ISMs) for urban communities. Here, a central monitoring and control station captures the energy generation/demand information of each microgrid and analyzes the availability/requirement, thereby executing the energy transactions among these ISMs. Such local energy exchanges among the ISMs reduce the issues with uncertain renewable energy and the dependency on the utility grid. To establish such useful ISMs, a well-established communication mechanism has to be adopted. In this view, this paper first reviews various state-of-the-art developments related to smart grids and then provides extensive insights into communication standards and technologies, issues/challenges, and future research perspectives for ISM implementation. Finally, a discussion is presented on advanced wireless technology, called LoRa (Long Range), and a modern architecture using the LoRa technology to establish a communication network for ISMs is proposed.

show abstract

Section: Insights Into Communication Standards and Technologiesmentioning

confidence: 99%

Communication Technologies for Interoperable Smart Microgrids in Urban Energy Community: A Broad Review of the State of the Art, Challenges, and Research Perspectives

Reddy

Kumar

Chakravarthi

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Mobile vehicles offload their computing tasks to cloud servers thousands of miles away via fiber optics and core networks [23]. Therefore, the upload time for transferring input data from the RSU to the cloud server must be considered.…”

Section: Cloud Server Computing Modelmentioning

confidence: 99%

Mobile Edge Computing Task Offloading Strategy Based on Parking Cooperation in the Internet of Vehicles

Shen

Chang

Niu

2022

Sensors

View full text Add to dashboard Cite

Due to the limited computing capacity of onboard devices, they can no longer meet a large number of computing requirements. Therefore, mobile edge computing (MEC) provides more computing and storage capabilities for vehicles. Inspired by a large number of roadside parking vehicles, this paper takes the roadside parking vehicles with idle computing resources as the task offloading platform and proposes a mobile edge computing task offloading strategy based on roadside parking cooperation. The resource sharing and mutual utilization among roadside vehicles, roadside units (RSU), and cloud servers (cloud servers) were established, and the collaborative offloading problem of computing tasks was transformed into a constraint problem. The hybrid genetic algorithm (HHGA) with a mountain-climbing operator was used to solve the multi-constraint problem, to reduce the delay and energy consumption of computing tasks. The simulation results show that when the number of tasks is 25, the delay and energy consumption of the HHGA algorithm is improved by 24.1% and 11.9%, respectively, compared with Tradition. When the task size is 1.0 MB, the HHGA algorithm reduces the system overhead by 7.9% compared with Tradition. Therefore, the proposed scheme can effectively reduce the total system cost during task offloading.

show abstract

“…In the offloading system, the offloading decision of mobile devices and the allocation of computing resources directly affect the time delay and energy consumption. 15 Offloading decision optimization can determine whether the mobile device is offloaded, while the optimization of computing resources determines how many resources are allocated. The coupling optimization of offloading decision and resource allocation belongs to the NP-hard problem.…”

Section: Introductionmentioning

confidence: 99%

A novel differential dynamic gradient descent optimization algorithm for resource allocation and offloading in the COMEC system

Chen

et al. 2022

Int J of Intelligent Sys

Self Cite

View full text Add to dashboard Cite

The multiuser cooperative offloading mobile edge computing (COMEC) system has attracted much attention because it can realize delay-sensitive tasks. However, in the coupling optimization of offloading decision and resource allocation, the existing numerical optimization algorithms are difficult to obtain high-quality optimization solutions. In this paper, we propose a differential dynamic gradient descent (DDGD) optimization algorithm to solve the above optimization problems. DDGD algorithm decomposes the constrained NP-hard optimization problem into two network layers and integrates the constraint function into a larger end-to-end training network. These two-layer networks encode the dependencies and optimization constraints between parameter hidden states, which cannot be captured by a numerical optimization model or a full connection layer neural network. Because the ring learning and self-repeating learning architecture are adopted and the information is stored in the differential dynamics network, the proposed algorithm can achieve better and more intelligent decision-making in searching the solution trajectory without setting accurate parameters in advance and

show abstract

Deep reinforcement learning for computation offloading in mobile edge computing environment

Cited by 62 publications

References 31 publications

Communication Technologies for Interoperable Smart Microgrids in Urban Energy Community: A Broad Review of the State of the Art, Challenges, and Research Perspectives

Communication Technologies for Interoperable Smart Microgrids in Urban Energy Community: A Broad Review of the State of the Art, Challenges, and Research Perspectives

Mobile Edge Computing Task Offloading Strategy Based on Parking Cooperation in the Internet of Vehicles

A novel differential dynamic gradient descent optimization algorithm for resource allocation and offloading in the COMEC system

Contact Info

Product

Resources

About