Original scientific paper

https://doi.org/10.17559/TV-20250505002644

Smart Forecasting Framework for Renewable Energy Systems Using Reinforcement Learning and Game Theory

Nageswari Devana ; Department of EEE, R. M. K. College of Engineering and Technology, Puduvoyal 601 206, India *
Marwa Obayya ; Department of Biomedical Engineering, College of Engineering, Princess Nourah bint Abdulrahman University, P. O. Box 84428, Riyadh 11671, Saudi Arabia
Menwa Alshammeri ; Department of Computer Science, College of Computer and Information Sciences, Jouf University, Saudi Arabia
Abdulwhab Alkharashi ; Department of Computer Science, College of Computing and Informatics, Saudi Electronic University, Saudi Arabia

* Corresponding author.

Abstract

This research introduces an advanced hybrid power forecasting framework for renewable energy systems, integrating Reinforcement Learning (RL) with Game Theory to address the limitations of traditional prediction models. Existing approaches often struggle with adaptability, scalability, and seamless integration across diverse and dynamic energy sources. To overcome these challenges, the proposed model leverages RL to adapt to fluctuating energy patterns in real time dynamically. At the same time, Game Theory is utilized to model strategic interactions and interdependencies among system entities. This combination fosters intelligent decision-making and cooperative behavior within the energy network. The model achieves a forecasting accuracy exceeding 98%, supported by an Average Learning Rate (ALR) of 0.85, demonstrating fast adaptation to changing conditions and reduced response latency. Furthermore, Game Theory facilitates collaborative system behavior, resulting in a high System Collaboration Index (SCI) of 0.92 -indicative of strong synergy among distributed components. Real-time feedback mechanisms enhance predictive reliability and improve energy inventory management by reducing discrepancies between demand and generation. Economically, the model contributes to a 20% improvement in Economic Efficiency (EE), reducing operational costs without sacrificing accuracy. By bridging technical forecasting precision with economic performance, this framework supports the development of an intelligent, cooperative, and resilient energy ecosystem. The proposed hybrid methodology offers a promising solution for sustainable energy management by improving forecast accuracy, system adaptability, and collaborative decision-making in renewable energy networks.

Keywords

economic efficiency; game theory; power forecasting; reinforcement learning; renewable energy; system collaboration

Hrčak ID:

353102

URI

https://hrcak.srce.hr/353102

Publication date:

7.11.2026.

Visits: 0 *