Orateur
Francis Durand
(Université Sorbonne Paris Nord)
Description
This work addresses the challenges of applying Stochastic Dual Dynamic Programming (SDDP) to infinite-horizon hydroelectric water management problems with continuous state and control spaces. While SDDP has proven effective in finite-horizon settings, its extension to the infinite-horizon case with a discount factor close to one introduces numerical difficulties when the discount rate is close to 1. To improve convergence, we propose a modified SDDP algorithm inspired by relative value iteration, incorporating additive shifts to accelerate the relevance of generated cuts. Numerical experiments highlight the practical benefits of our approach for long-term planning.
Authors
Bernardo Freitas Paulo da Costa
(Fundação Getulio Vargas)
Francis Durand
(Université Sorbonne Paris Nord)
M.
Mathis Azéma
(CERMICS)
Vincent Leclere
(ENPC)
