Survey maps a unified AI framework for renewable-dominated power grids

A new survey published in Artificial Intelligence Review argues that the intelligence behind renewable-dominated power systems has been built in fragments, and that the field needs a single framework connecting forecasting, explainability and optimization. The paper, authored by Tariq Ali, Muzna Sarwar, Husam S. Samkari, Farrukh Jamal, Mohammad Hijji, Muhammad Ayaz, Mohammed F. Allehyani, Muhammad Irfan and eL-Hadi M. Aggoune, was published open access on 21 October 2026 as a peer-reviewed, accepted article that is citable and carries a permanent DOI.

The authors frame the problem around the operational realities of grids increasingly supplied by wind and solar. The transition toward renewable-dominant power systems has introduced significant operational challenges, they write, because of increased uncertainty, reduced system inertia and the proliferation of distributed energy resources. Artificial intelligence techniques have been widely adopted for renewable energy forecasting and grid optimization, but existing studies primarily address these components in isolation, leading to what the survey calls fragmented system-level intelligence.

To address that fragmentation, the survey systematically integrates three dimensions of modern intelligent energy systems. The first is hybrid forecasting. The authors review advanced hybrid learning architectures, including decomposition-based models, ensemble frameworks, and deep learning-metaheuristic hybrids, highlighting their effectiveness in capturing nonlinear and stochastic renewable dynamics.

The second dimension is explainable artificial intelligence, or XAI. The survey examines XAI techniques that enhance transparency, interpretability and trust in AI-driven energy systems.

The third dimension is the evolution from centralized optimization to distributed collective intelligence. The authors analyze multi-agent reinforcement learning, swarm intelligence and federated learning paradigms as alternatives to a single central controller.

Unlike conventional surveys, the paper states, this work provides a unified system-level perspective by explicitly linking predictive uncertainty with decentralized decision-making and explainability mechanisms. A conceptual integration framework is presented to demonstrate how probabilistic forecasting, risk-aware multi-agent reinforcement learning policies and interpretable models can jointly enable robust and scalable grid operation. The framework is a conceptual proposal rather than a tested deployment; the survey’s contribution is the structured linkage of the three components, not an empirical demonstration on a live grid.

The survey also identifies key research challenges related to scalability, computational efficiency, uncertainty propagation and real-time deployment. These open problems indicate where the proposed integration still falls short of operational requirements: methods must scale to large networks, run within the computational budgets of grid operations, propagate uncertainty coherently from forecast to decision, and function under real-time constraints.

The authors position the work as a structured reference for researchers and practitioners, outlining future directions toward trustworthy, adaptive and cooperative AI-driven energy systems. As a survey, it synthesizes and organizes existing research rather than reporting new experimental results, and its claims about the effectiveness of hybrid architectures reflect the literature it reviews.

Subject of Research: A survey integrating hybrid forecasting, explainable AI and multi-agent optimization for renewable-dominated power systems

Article Title: Toward unified intelligence in renewable-dominant power systems: a survey of hybrid forecasting, explainable AI, and multi-agent optimization

Article References: Ali, T., Sarwar, M., Samkari, H. S., Jamal, F., Hijji, M., Ayaz, M., Allehyani, M. F., Irfan, M., & Aggoune, E.-H. M. (2026). Toward unified intelligence in renewable-dominant power systems: a survey of hybrid forecasting, explainable AI, and multi-agent optimization. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11711-6

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11711-6

Keywords: renewable energy, power systems, artificial intelligence, hybrid forecasting, explainable AI, multi-agent reinforcement learning, federated learning, grid optimization, survey, distributed energy resources, Toward, unified