Lightweight Federated Learning with Firefly Optimization for Pollution-Aware Internet of Green Things
Journal of Quantum Nano- Green Environmental Systems · 2026
DOI: 10.70023/21602 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Quantum Nano- Green Environmental Systems, 2026. Verfasst von Abdullah Khatib und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Rising pollution and environmental issues have further fueled the need for smart, energy-efficient approaches to the Pollution-Aware Internet of Green Things (IoGT). Traditional, centralized approaches are not scalable on resource-limited devices, incur high communication overhead, and lack privacy. A lightweight federated learning incorporating Firefly Optimization (LFL-FO) is proposed for distributed air quality prediction using the India-multi-city-air-quality-dataset, which features air quality measurements (PM2.5, PM10, NO₂, SO₂, CO, and O₃) from geographically distributed sensors. Each node trains a model on its own data and then securely sends a consensus of the model parameters to the central server without transferring raw information. Firefly Optimization adjusts learning parameters and client participation to achieve faster convergence and reduce computational complexity. The experiments conducted on the India-multi-city-air-quality-dataset yielded an average F1 score of 98.5%, an average communication cost of 16.8 MB, an average energy consumption of 2.3 J, a model size of 5.6 MB, and an average inference latency of 28 ms. Introducing lightweight functionality combined with Fireflyinspired optimizations of pollution monitoring in smart cities and sustainable environmental management systems while preserving end-user privacy.
Zitieren
APA
Khatib, A., Mansoori, M., & Kumar, R. (2026). Lightweight Federated Learning with Firefly Optimization for Pollution-Aware Internet of Green Things. Journal of Quantum Nano- Green Environmental Systems. https://doi.org/10.70023/21602
Vancouver
Khatib A, Mansoori M, Kumar R. Lightweight Federated Learning with Firefly Optimization for Pollution-Aware Internet of Green Things. Journal of Quantum Nano- Green Environmental Systems. 2026. doi:10.70023/21602
BibTeX
@article{Khatib2026,
author = {Khatib, Abdullah and Mansoori, Mariam and Kumar, Rahul},
title = {Lightweight Federated Learning with Firefly Optimization for Pollution-Aware Internet of Green Things},
journal = {Journal of Quantum Nano- Green Environmental Systems},
year = {2026},
doi = {10.70023/21602}
}
EUVIMED