Work

Energy aware Multi Agent Deep Queue Optimization for efficient Resource Allocation and Task Offloading in Cloud Edge

Year: 2024
Type: article
Abstract: IoT devices finds its application in almost all the fields leading to streaming of enormous amount of big and small data. This leads to many operational problems in load balancing, energy management, latency in processing and storage methods. Edge and cloud computing is leveraged as a potential solution to resolve these problems through its resourceful architectures and on demand services. However, achieving optimal energy efficiency and latency in time critical medical applications is still an open-ended research topic, that draws the attention of the researchers. This work proposes Energy aware Multi Agent Deep Queue Optimisation (E-MADQ) technique that classifies the priorities of the medical tasks using Ensemble Empirical Mode Decomposition (EEMD) smoothened with Extreme Gradient Boost (XGBoost) by extracting Intrinsic Mode Functions and averaging their spectrum characteristics. The offloading decision is made using the multi-agent deep queue optimization where the rewards are calculated based on the energy level of the IoT devices, which is very crucial parameter. By this, the tasks that demands high attention will not be left in starvation and proper resources allocation is done for time critical tasks. The experimental simulation of the proposed methodology shows that a good improvement in service parameters such as mean delivery time, communication and computing delay, execution time and energy level can be attained. In future, the classification of task priorities can be done with powerful deep learning techniques with more focus on dynamism.
Language: English
Open Access status: diamond
OpenAlex ID: W7218180013
FWCI:
Cites: 20
Cited by:
Related to: 10