HyperP2P: Bipartite Hypergraph Matchings for Peer-to-Peer Energy Sharing

Research Project, 2022

Recent research advocates Peer-to-Peer (P2P) energy sharing as a way to make the most of local energy resources, reducing the bill of each participating end-user. However, setting up such systems in practice is made very difficult by the computational difficulties and network burden posed by computing the peer matching in an online fashion (based on the peers' local historical data and some forecast of the future), the large-scale size of potential applications (thousands to hundred of thousands of possible end-users for city-wide adoption of such sharing economy) and the peers' privacy concerns (advocating sharing the least amount of data with the least number of entities for the maximum monetary benefits). Further practical and theoretical tools are thus needed to tackle the complex computational problems that stem from the integration of large-scale distributed renewable energy resources. In this context, we propose innovative ICT solutions capable of advising a very effective matching of peers using only little computing resources and data (eg some aggregated statistics). To prove the effectiveness of our approach, we develop a mathematical framework to show that our heuristic solutions are indeed effective. The potential outcomes are twofold: (1) linking the matching problem to known and already studied settings allows us to re-use algorithms and results from the literature and (2) our approach provides new formal guarantees of effectiveness for the proposed practical algorithms that may enhance their deployment.