A Parallel Solution to Large Scale Hydropower Scheduling - By reducing the number of synchronization points in the Stochastic Dual Dynamic Programming (SDDP) Algorithm

Abstract

Stochastic dual dynamic programming (SDDP) has become a popular algorithm used in practical long-term scheduling of hydro power systems. The SDDP algorithm is significantly more computationally demanding than most heuristic-based scheduling methods, but can be designed to take advantage of parallel processing. This thesis presents a novel parallel scheme for the SDDP algorithm, where the stage-wise synchronization point traditionally used in the backward iteration of the SDDP algorithm is either partially or fully relaxed. The proposed scheme was tested on a realistic model of a Norwegian water course, proving that the partial synchronization point relaxation significantly improves parallel efficiency.