ºÚÁÏÕýÄÜÁ¿

Electricity markets have been around for a long time (i.e., since the early 1990s) in countries like the UK. Their context has evolved tremendously since then, with increased penetration from renewable energy sources, decentralisation and digitisation of energy systems, and new computing capabilities (e.g., with AI). There is consensus that it is time to rethink electricity markets, with potential changes that could go from incremental to (possibly more likely) fundamental. In parallel, AI has been widely seen as an enabler to the operation of future power/energy systems, though very little consideration has been given to how it may get to the core of electricity markets. With this talk, I will motivate why I think it is the right moment to ask ourselves this question, and will discuss how I think this will happen. The main idea is to replace the core of electricity markets (i.e., the optimisation problem formulation and the solvers employed) by an AI. This may allow to rethink what electricity markets are for, the way agents are coordinated, pricing, etc. We will then discuss how we can build on the work of different groups over the world who are looking into AI-based approaches to solve optimal power flow problems and similar. We will subsequently zoom on approaches dedicated to solving linear programs with AI, and prospects to solve unit commitment problems with AI. The talk will close with a proposal roadmap to get AI at the core of future electricity markets.

Pierre Pinson received both MSc and PhD degrees in France, with focus on applied maths and applications to energy systems. He spent most of his career in Denmark, at DTU, before to move to the UK a few years ago. He has held academics positions in Statistics, Electrical Engineering and Management Science. He is now a Professor at Imperial College London, Dyson school, and Deputy-Head of the School. He is also a Chief Scientist at Halfspace, which was recently acquired by Accenture. He has honorary appointments at DTU and Aarhus University (Denmark). He is the Editor-in-Chief of the International Journal of Forecasting, the leading journal in its field. His works are mainly focused on forecasting, optimization and game theory, for power and energy systems applications. He is an IEEE Fellow and an IIF Director. In 2024, he led the Molslinjen/Halfspace team that received the INFORMS Franz Edelman Award.