Fog-GMFA-DRL: Enhanced deep reinforcement learning with hybrid grey wolf and modified moth flame optimization to enhance the load balancing in the fog-IoT environment

Gupta, Shally; Singh, Nanhay

doi:10.1016/j.advengsoft.2022.103295

Cited by 6 publications

(4 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With each iteration of the method, the fog node's network experience grows, and over time it learns to assign the sub-task to a node with a lighter workload and faster processing speed. Contrary to the proposed way, alternative solutions (such as those in sources [6,8]) conduct the load balancing algorithm before adding overhead to the fog node, which degrades the performance of the aforementioned systems and lengthens their latency. Another benefit is that, in the comparative methods for allocating subtasks to surrounding nodes, it is only necessary to be aware of this node's position and capacity, which may be determined by making a few numbers of laborious computations.…”

Section: Evaluation and Simulationmentioning

confidence: 99%

“…A number of current load balancing techniques have been compared to the performance of the suggested method, and in this simulation, random and proportional SALB load balancing techniques have been employed as benchmarks [6,8]. The SALB approach examines the adjacent nodes' capacities after the fog node is overloaded and delivers the subtask to the node with the highest capacity and at least 40% of its capacity.…”

Section: Evaluation and Simulationmentioning

confidence: 99%

“…Two distinct forms of load balancing are utilized in dispersed environments: static load balancing and dynamic load balancing. When choosing a load balancer, static load balancing ignores the target fog node's state [6]. On the other hand, dynamic load balancing chooses which fog node to route traffic based on its present status [7].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improving Load Balance in Fog Nodes by Reinforcement Learning Algorithm

DING,

ZHANG

2024

IJACSA

View full text Add to dashboard Cite

Fog computing is a distributed computing concept that brings cloud services out to the network's edge. Real-time user queries and data streams are processed by cloud nodes. Tasks should be evenly divided among fog nodes in order to maximize speed and efficiency, optimize resource efficiency, and reaction time. Real-time user requests and data flow processing are done by cloud nodes. Nodes in a network must share responsibilities in a balanced manner in order to maximize speed and efficiency, resource efficiency, and reaction time, hence in this article, a novel approach is presented. When it comes to fog computing, load balancing essential suggested to be improved. According to the suggested algorithm, a task submitted to the fog node via a mobile device would be processed by the fog node using reinforcement learning before being passed on to another fog node. Neighbor or let the cloud handle it. According to the simulation findings, the suggested algorithm has achieved a reduced execution time than other compared approaches by properly allocating the work among the nodes. Consequently, the suggested technique has reduced the chance of incorrect job assignment by 24.02% and the response time to the user by 31.60% when compared to similar methods.

show abstract

Section: Evaluation and Simulationmentioning

confidence: 99%

Section: Evaluation and Simulationmentioning

confidence: 99%

See 1 more Smart Citation

Improving Load Balance in Fog Nodes by Reinforcement Learning Algorithm

DING,

ZHANG

2024

IJACSA

View full text Add to dashboard Cite

show abstract

“…By combining Grey Wolf Optimisation (GWO) and the Modified Moth Flame algorithm (MMFA), the unique hybrid algorithm proposed by (Gupta and Singh, 2022) enhances Deep reinforcement learning (DRL) by furnishing a local search mechanism for scheduling tasks in fog computing scenarios. The authors offer an alternative fitness function that factors in task dependencies, resource availability, and communication costs.…”

Section: Literature Surveymentioning

confidence: 99%

AGWO: Cost Aware Task Scheduling in Cloud Fog Environment Using Hybrid Metaheuristic Algorithm

Medishetty,

2023

IJERR

View full text Add to dashboard Cite

In IoT concepts, efficient approaches like cloud-fog computing are emerging, enhancing system benefits. The performance and output of such frameworks can be greatly improved by optimized scheduling of Internet of Things (IoT) task requests. This study presents a novel technique for scheduling Internet of Things requests in a cloud-fog environment, based on an adaption of ant grey wolf optimisation (AGWO). By combining the operators of ant colony optimisation (ACO) and grey wolf optimisation (GWO), AGWO aims to improve the speed and quality of ACO's solution discovery. The suggested AGWO approach is evaluated using numerous datasets of varying sizes, both synthetic and real-world. The effectiveness of AGWO is further investigated by comparing it to standard metaheuristic methods. The experimental results demonstrate that AGWO is superior to competing strategies in terms of makespan time and cost while resolving the task scheduling problem.

show abstract