Open Access

ARTICLE

Deep Learning and Heuristic Optimization for Secure and Efficient Energy Management in Smart Communities

Murad Khan^1,*, Mohammed Faisal¹, Fahad R. Albogamy², Muhammad Diyan³

1 Department of Computer Science and Engineering, Kuwait College of Science and Technology, Doha District, Kuwait City, P.O. Box 35001, Kuwait
2 Computer Sciences Program, Department of Mathematics, Turabah University College, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia
3 Department of Computing & Games, Teesside University, Middlesbrough, TS1 3BX, UK

* Corresponding Author: Murad Khan. Email:

(This article belongs to the Special Issue: Emerging Technologies in Information Security )

Computer Modeling in Engineering & Sciences 2025, 143(2), 2027-2052. https://doi.org/10.32604/cmes.2025.063764

Received 23 January 2025; Accepted 08 April 2025; Issue published 30 May 2025

Abstract

The rapid advancements in distributed generation technologies, the widespread adoption of distributed energy resources, and the integration of 5G technology have spurred sharing economy businesses within the electricity sector. Revolutionary technologies such as blockchain, 5G connectivity, and Internet of Things (IoT) devices have facilitated peer-to-peer distribution and real-time response to fluctuations in supply and demand. Nevertheless, sharing electricity within a smart community presents numerous challenges, including intricate design considerations, equitable allocation, and accurate forecasting due to the lack of well-organized temporal parameters. To address these challenges, this proposed system is focused on sharing extra electricity within the smart community. The working of the proposed system is composed of five main phases. In phase 1, we develop a model to forecast the energy consumption of the appliances using the Long Short-Term Memory (LSTM) integrated with the attention module. In phase 2, based on the predicted energy consumption, we designed a smart scheduler with attention-induced Genetic Algorithm (GA) to schedule the appliances to reduce energy consumption. In phase 3, a dynamic Feed-in Tariff (dFIT) algorithm makes real-time tariff adjustments using LSTM for demand prediction and SHapley Additive exPlanations (SHAP) values to improve model transparency. In phase 4, the energy saved from solar systems and smart scheduling is shared with the community grid. Finally, in phase 5, SDP security ensures the integrity and confidentiality of shared energy data. To evaluate the performance of energy sharing and scheduling for houses with and without solar support, we simulated the above phases using data obtained from the energy consumption of 17 household appliances in our IoT laboratory. Finally, the simulation results show that the proposed scheme reduces energy consumption and ensures secure and efficient distribution with peers, promoting a more sustainable energy management and resilient smart community.

---

Keywords

---