In February 2020, I arrived as a postdoctoral reaseacher in the GERAD center. I am working under the supervision of Sébastien Le Digabel and in partnership with Hydro-Québec (IREQ).

My research interests are about distributed optimization and game theory and multi-agents systems, and the applications of these methods to the decentralized management of electric systems. These topics includes efficiency and equilibria analysis, Smart Grid, renewable energies, demand response management and privacy concerns.

I finished a PhD, done in a partnership between EDF Lab and the Inria team TROPICAL located in the CMAP at École polytechnique. This Ph.D. thesis was supervised by Stéphane Gaubert, head of TROPICAL team, and by both Nadia Oudjane and Olivier Beaude at EDF Lab, in the department OSIRIS “Optimization, SImulation, RIsk and Statistics” located in Saclay, France.

I defended my PhD Game theory and Optimization Methods for Decentralized Electric Systems on December 5, 2019 at the Ecole polytechnique.

Visit my website : http://www.cmap.polytechnique.fr/~paulin.jacquot/

The project will be conducted in partnership with Hydro-Québec ({IREQ} research institute), a major contributor of IVADO. In addition to the academic supervision of Sébastien Le Digabel (professor at the Department of Mathematics and Industrial Engineering, Polytechnique Montréal, Canada, and member of the GERAD research center), the project will benefit from interactions with IREQ researchers Laurent Lenoir (Scientific Leader in T\&D Grid Operation, Hydro-Québec, IREQ) and Stéphane Alarie (IREQ and GERAD associate member).

In order to satisfy the 2030 and 2050 CO2 reduction objectives in Québec, a massive electrification of usages, in particular in transport (electric vehicles) and heating systems, is necessary (cf Report Dunsky in 2019). To cope with this large increase of the demand, the electricity system has to be adapted, starting with the development of renewable energy sources. The transmission and distribution grids will be subject to heavier stress and constraints, and their management and optimization call for advanced methods to be developed.

Distributed Energy Resources (DERs) such as storage, electric vehicles, consumption flexibilities, local renewable production and microgrids, will be an essential component of this future electricity system.

The decentralized optimization of DERs is considered a key tool for the integration of local renewable production and for the efficiency of the grid. The management of DERs can be used for real-time control of the grid state and constraints, thus enhancing the system resilience while limiting the need for costly infrastructure investments.

From an optimization perspective, this calls for the development of new decentralized algorithms and methods (e.g. \emph{Demand Response} in the smart grid literature \cite{PaulinTSG17}).

The project aims to study the interactions between the multiple actors of the grid (system operator, producers, consumers, etc) and the impacts on the grid constraints, in particular at the transmission level. This will require advanced mathematical techniques from operations research, continuous optimization, game theory and data analysis. In particular, this project addresses the following questions:

• establish and study a model of management and control of the future electricity system, involving multiple distributed actors and presence of DERs. Possible optimization models could consider \emph{multi-level} optimization or \textit{multi-level game} frameworks (e.g. \emph{Stackelberg games}) \cite{maharjan2013dependable}, \cite{yu2015real};
• develop distributed algorithms for the real-time control of the grid: starting from the grid current state and available controlable DERs, compute and implement optimal decisions through ad-hoc coordination. The methods have to be in line with the hierarchical structure of the grid (transmission system operator and substations) and will consider an AC optimal power flow model (OPF) for a realistic model of the grid constraints (see recent results on convex relaxations of the OPF problem \cite{bingane2018tight}, \cite{peng2016distributed}). Derivative free and black-box optimization techniques could be used to adress large scale OPF problems while capturing engineering details, as suggested in \cite{bienstock2019variance};
• analyze and use data sets provided by Hydro-Québec (historical data of load profiles and transmission network states and parameters), enabling to perform realistic simulations;
• quantify the benefits of the exploitation of consumers flexibilities related to the savings on the network infrastructure investments, in longer term perspective. A possible approach relies on bilevel optimization model \cite{asensio2017bi}.

avancement-CRs réunions