Scenario tree reduction and operator decomposition method for stochastic optimization of energy systems

The drop in production costs of distributed energy production systems and electrochemical storage systems coupled with the evolution of regulations make it possible to build local energy management operations. Development of such operations will be all the easier if the management systems allow the various involved players to reduce their electricity bills and/or their greenhouse gas emissions To this end, the various energy systems need to be optimized, a challenging task for mainly two reasons: First, the random nature of consumption/production. Second, the complex mathematical model needed to accurately represent the physics and practical constraints of the energy systems.
In this context, IFPEN develops a high-performance solution handling the two major challenges: stochasticity management and model complexity. In this context the Ph.D. candidate will work with state-of-the-art methodologies in stochastic programming and develop original methods for reducing scenario trees. To do this, he or she will use tools  from optimal transport and data science to determine the best structure and the minimum size of the scenario tree to achieve, in combination with splitting methods in optimization, good computational performance, and solution accuracy.

Keywords: Stochastic Optimization, data science, optimal control, scenario trees

• Academic supervisor    Dr. de OLIVEIRA Welington (www.oliveira.mat.br) Center of applied mathematics MINES-ParisTech (www.cma.mines-paristech.fr)  
• Doctoral School    STIC – Sciences et technologies de l’information et de la communications
• IFPEN supervisor     Dr., MALISANI Paul, Mathématiques appliquées,  paul.malisani@ifpen.fr  
• PhD location    IFP Energies nouvelles, Rueil-Malmaison, France
• Duration and start date    3 years, starting in fourth quarter 2022
• Employer    IFP Energies nouvelles, Rueil-Malmaison, France
• Academic requirements    Master degree 2 in applied mathematics, optimization, data science,
• Language requirements    English essential, french desirable
• Other requirements     Python, LaTeX