Mustapha Muhammad Muhammad¹*, Bukar Umar Musa2, Jafaru Usman3, Ibrahim Mustapha4and Micheal Abimbola Oladosu5
¹,2,3,4 Department of Electrical Engineering, University of Maiduguri, Maiduguri, Borno State, Nigeria.
5Department of Research and Linkages, Research Hub Nexus Institute, Ayobo, Lagos, Nigeria
(✉) Corresponding Author:
Received: Apr 24, 2026/ Revised: May 28 2026 /Accepted: June 10, 2026
Highlights
- Systematically reviews Federated Learning, Game Theory, and Explainable AI in decentralized energy systems.
- Examines privacy preservation, fairness, and transparency in microgrids and P2P energy trading.
- Identifies the lack of unified FL–GT–XAI frameworks for trustworthy energy management.
- Highlights challenges in privacy-aware pricing, incentive fairness, and interpretable decision-making.
- Provides directions for resilient, equitable, and transparent decentralized energy markets.
Abstract
The transition toward decentralized, prosumer-driven energy systems has intensified the need for privacy-preserving analytics, fair market mechanisms, and transparent decision-making. While Federated Learning (FL), Game Theory (GT), and Explainable Artificial Intelligence (XAI) independently address these challenges, their integration within decentralized energy systems remains fragmented. This paper presents a systematic review of 52 selected peer-reviewed studies, supported by additional foundational references, published between 2020 and 2026. The review examines the applications of Federated Learning (FL), Game Theory (GT), and Explainable Artificial Intelligence (XAI) in microgrids, peer-to-peer (P2P) energy trading, and decentralized energy management systems. Using a structured methodology comprising comprehensive database searches, clearly defined inclusion and exclusion criteria, and cross-paradigm thematic synthesis, the review identifies critical gaps in the current literature. The findings reveal that existing research largely treats FL, GT, and XAI as isolated paradigms, with no study providing a unified tri-layer framework integrating privacy-preserving learning, strategic market design, and interpretable decision-making. Furthermore, most contributions rely on idealized simulations and originate from technologically advanced regions, offering limited applicability to fragile, data-scarce environments such as conflict-affected Northeastern Nigeria. Key limitations identified include insufficient support for privacy-aware market pricing, limited fairness guarantees in strategic interactions, and weak adoption of explainability in decentralized decision systems. This review represents one of the first consolidated assessments of FL–GT–XAI convergence in decentralized energy systems and contextualizes these paradigms within the operational realities of emerging and fragile regions. The synthesis highlights the urgent need for integrated architectures capable of simultaneously ensuring privacy preservation, incentive compatibility, and transparent decision support for institutional microgrids. Insights from this study provide a foundation for the development of trustworthy, equitable, and resilient decentralized energy markets.
Keywords: Decentralized energy systems, Federated learning, Game theory, Explainable AI, Peer-to-peer energy trading, Microgrids, Privacy, Energy equity.
How to cite this article
Muhammadi, M. M., Musa, B. U., Usman, J., Mustapha, I., & Oladosu, M. A. (2026). Privacy, fairness, and transparency in decentralized energy systems: A systematic review of FL–GT–XAI approaches. Science Archives, 7(2), 240–259. https://doi.org/10.47587/SA.2026.7222
This work is licensed under a Creative Commons Attribution 4.0 International License













