AI optimization for peer-to-peer energy sharing: Trends, methods, and future directions
Date
2026
Authors
Filipovich, Daniil
Journal Title
Journal ISSN
Volume Title
Publisher
University of Victoria
Abstract
As clean energy technology advances, Peer-to-peer (P2P) energy sharing has transitioned from a theoretical concept to an emerging real-world solution. A meta-analysis of over 25 academic sources from 2020 to 2025 was performed to track the evolution of P2P modelling. Early models (circa 2020) relied on predefined, fixed variables (such as weather, energy use, and energy storage) within centralized decision-making frameworks. While providing control, these models lacked the flexibility required for volatile real-world conditions. Over the past 5 years, the literature has shifted toward decentralized environments, accounting for changing variables and outcome probabilities. This shift has been accelerated by the growing availability of empirical data from communities that began peer-to-peer adoption, in Europe, the USA, and Australia, among others. According to industry trends, the next phase in this evolution is using real-world data for AI-driven data optimization. By requesting and compiling such data and replacing existing synthetic inputs and heuristic assumptions with empirical data, future models may help bridge the gap between existing theoretical modelling and real-world deployment, enabling real-time optimization of pricing, energy load distribution, and storage management.
Description
Keywords
P2P energy sharing, decentralized energy management, stochastic optimization, artificial intelligence, empirical data analysis, grid flexibility, Jamie Cassels Undergraduate Research Awards (JCURA)