EUVIMEDEUVIMED entdecken
Crossref

A Privacy - Preserving Federated Learning Framework for Dual-Resource Sustainability Optimization

Dr. Arikumar K S, Jayeesh Hari Varma V, Naveena Sahitya Kothapalli, Tanisha M

International Journal of Science, Strategic Management and Technology · 07.05.2026 · 02 (05) · 1-9

DOI: 10.55041/ijsmt.v2i4.540 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Science, Strategic Management and Technology, online erschienen am 07.05.2026. Verfasst von Dr. Arikumar K S und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Centralized machine learning for smart home energy management has run into problems because household consumption data is very private. To solve this problem, we present GreenFed, a federated learning (FL) framework that respects privacy and optimizes water and electricity use in homes without ever showing raw sensor readings. GreenFed uses Long Short-Term Memory (LSTM) networks on 101 simulated households taken from the CEEW Bareilly BR02 dataset. It uses the FedAvg protocol to coordinate weight updates over ten communication rounds. A new composite metric called the Dual GreenScore (from 0 to 100) combines signals for water and electricity efficiency into a single, easy-to-read sustainability index. The RMSE for predicting electricity went down by 4.7 percent (from 0.007311 to 0.006966) and the RMSE for predicting water went down by 10.5 percent (from 0.011535 to 0.010318). A full-stack dashboard built with React shows GreenScore in real time, lets you interact with a behavioral simulator, analyzes your CO2 footprint, and sends you automated sustainability reports. CO2 emissions are computed using India’s BEE-specified grid carbon intensity of 0.82 kg/kWh. To our knowledge, no prior FL framework has concurrently addressed electricity and water optimization under a unified privacy-compliant sustainability index grounded in Indian smart meter data — a gap that GreenFed directly fills. The paper is organized as follows: Section II surveys prior literature. Section III details GreenFed’s architecture and methodology. Experimental outcomes are presented in Section IV, followed by implementation specifics in Section V. Section VI concludes with future directions.

Zitieren

APA

S, D. A. K., V, J. H. V., Kothapalli, N. S., & M, T. (2026). A Privacy - Preserving Federated Learning Framework for Dual-Resource Sustainability Optimization. International Journal of Science, Strategic Management and Technology, 02(05), 1-9. https://doi.org/10.55041/ijsmt.v2i4.540

Vancouver

S DAK, V JHV, Kothapalli NS, M T. A Privacy - Preserving Federated Learning Framework for Dual-Resource Sustainability Optimization. International Journal of Science, Strategic Management and Technology. 2026;02(05):1-9. doi:10.55041/ijsmt.v2i4.540

BibTeX
@article{S2026,
  author = {S, Dr. Arikumar K and V, Jayeesh Hari Varma and Kothapalli, Naveena Sahitya and M, Tanisha},
  title = {A Privacy - Preserving Federated Learning Framework for Dual-Resource Sustainability Optimization},
  journal = {International Journal of Science, Strategic Management and Technology},
  year = {2026},
  volume = {02},
  number = {05},
  pages = {1--9},
  doi = {10.55041/ijsmt.v2i4.540}
}
Kostenlos starten →