Real-Time Task Assignment Approach Leveraging Reinforcement Learning with Evolution Strategies for Long-Term Latency Minimization in Fog Computing

Mai, Long; Dao, Nhu‐Ngoc; Park, Min‐Ho

doi:10.3390/s18092830

Cited by 38 publications

(32 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep reinforcement learning has also been used by the work of Mai et al for Real‐Time Task Assignment (RTTA) using a neural network, trained by reinforcement learning approach using evolutionary strategies for scheduling real‐time jobs.…”

Section: Related Workmentioning

confidence: 99%

A job scheduling algorithm for delay and performance optimization in fog computing

Jamil

Shojafar

Ahmed

et al. 2019

Concurrency and Computation

107

View full text Add to dashboard Cite

Summary Due to an ever‐increasing number of Internet of Everything (IoE) devices, massive amounts of data are produced daily. Cloud computing offers storage, processing, and analysis services for handling of such large quantities of data. The increased latency and bandwidth consumption is not acceptable to real‐time applications like online gaming, smart health, video surveillance, etc. Fog computing has emerged to overcome the increase in latency and bandwidth consumption in Cloud computing. Fog Computing provides storage, processing, networking, and analytical services at the edge of a network. As Fog Computing is still in its infancy, its significant challenges include resource‐allocation and job‐scheduling. The Fog devices at the edge of the network are resource‐constrained. Therefore, it is important to decide the assignment and scheduling of a job on a Fog node. An efficient job scheduling algorithm can reduce energy consumption and response time of an application request. In this paper, we propose a novel Fog computing scheduler that supports service‐provisioning for Internet of Everything, which optimizes delay and network usage. We present a case study to optimally schedule the requests of Internet of Everything devices on Fog devices and efficiently address their demands on available resources on every Fog device. We consider delay and energy consumption as performance metrics and evaluate the proposed scheduling algorithm using iFogSim in comparison with existing approaches. The results show that the delay and network usage of the proposed scheduler improve by 32% and 16%, respectively, in comparison with FCFS approach.

show abstract

Section: Related Workmentioning

confidence: 99%

A job scheduling algorithm for delay and performance optimization in fog computing

Jamil

Shojafar

Ahmed

et al. 2019

Concurrency and Computation

107

View full text Add to dashboard Cite

show abstract

“…The drawback of the PIOTS algorithm is time and computing resource consumption for SOM operation. To address the real-time resource management in MEC, Mai et al [18] proposed a reinforcement learning (RL)-based approach to distribute IoT tasks among edge servers to reduce task execution latency. Although the proposed algorithm can provide online task decision, RL operation itself requires significant computing resource.…”

Section: B a Review Of Cutting-edge Approachesmentioning

confidence: 99%

Self-Calibrated Edge Computation for Unmodeled Time-Sensitive IoT Offloading Traffic

et al. 2020

Self Cite

View full text Add to dashboard Cite

With the characterizing benefits of ultra-low latency, contextual computing, and mobile scalability, mobile edge computing (MEC) is considered as a key enabler for realizing a tremendous boom in heterogeneously time-sensitive Internet-of-Things (IoT) services in the fifth-generation (5G) ecosystems. However, achieving low-latency comes at a cost of energy-efficiency reduction. To address and balance this tradeoff, this paper proposes a joint optimization of energy consumption and latency satisfaction in MEC servers, called latency-aware green (LAG) computing algorithm. To fully consider the heterogeneity of IoT services offloaded to the MEC servers, offloading traffic at the MEC servers is assumed to be unmodeled and unpredictable. Using the proposed LAG algorithm, each MEC server autonomously and dynamically calibrates its own computing frequency based on the current status of the workload buffer size and computational workload arrival rate. This dynamic calibration provides minimum energy consumption for the workload computation while maintaining the computational latency stabilized under a desired threshold. Evaluation results show that the proposed algorithm maintains stable MEC servers in an energyefficient manner. INDEX TERMS edge computing, Internet of Things, mobile offloading, unmodeled traffic VOLUME 8, 2020 This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication.

show abstract

“…Our work, on the other hand, considers a stochastic scenario with additional optimisation criteria and realistic pricing schemes. Moreover, given the benefits of applying RL to problems with large search spaces, such as the game of Go [6], previous work also applied RL to scheduling tasks to Edge computing [14]. RL has been used for solving deterministic and stochastic scheduling problems [14,16] and elasticity of DSP applications.…”

Section: Related Workmentioning

confidence: 99%

“…Moreover, given the benefits of applying RL to problems with large search spaces, such as the game of Go [6], previous work also applied RL to scheduling tasks to Edge computing [14]. RL has been used for solving deterministic and stochastic scheduling problems [14,16] and elasticity of DSP applications. Vengerov et al [28] show an RL framework for performing adaptive reconfiguration of dynamic resource allocations with Fuzzy.…”

Section: Related Workmentioning

confidence: 99%

Multi-Objective Reinforcement Learning for Reconfiguring Data Stream Analytics on Edge Computing

Veith

Souza

Assunção

et al. 2019

Proceedings of the 48th International Conference on Parallel Processing

View full text Add to dashboard Cite

There is increasing demand for handling massive amounts of data in a timely manner via Distributed Stream Processing (DSP). A DSP application is often structured as a directed graph whose vertices are operators that perform transformations over the incoming data and edges representing the data streams between operators. DSP applications are traditionally deployed on the Cloud in order to explore the virtually unlimited number of resources. Edge computing has emerged as a suitable paradigm for executing parts of DSP applications by offloading certain operators from the Cloud and placing them close to where the data is generated, hence minimising the overall time required to process data events (i.e., the end-toend latency). The operator reconfiguration consists of changing the initial placement by reassigning operators to different devices given target performance metrics. In this work, we model the operator reconfiguration as a Reinforcement Learning (RL) problem and define a multi-objective reward considering metrics regarding operator reconfiguration, and infrastructure and application improvement. Experimental results show that reconfiguration algorithms that minimise only end-to-end processing latency can have a substantial impact on WAN traffic and communication cost. The results also demonstrate that when reconfiguring operators, RL algorithms improve by over 50% the performance of the initial placement provided by state-of-the-art approaches.

show abstract

Real-Time Task Assignment Approach Leveraging Reinforcement Learning with Evolution Strategies for Long-Term Latency Minimization in Fog Computing

Cited by 38 publications

References 34 publications

A job scheduling algorithm for delay and performance optimization in fog computing

A job scheduling algorithm for delay and performance optimization in fog computing

Self-Calibrated Edge Computation for Unmodeled Time-Sensitive IoT Offloading Traffic

Multi-Objective Reinforcement Learning for Reconfiguring Data Stream Analytics on Edge Computing

Contact Info

Product

Resources

About