Aplicação de programação dinâmica estocástica e dual estocástica na otimização da operação do sistema de reservatórios de múltiplos usos do Rio São Francisco

Abstract

The objective of this work is to apply Stochastic Dynamic Programming and Dual Dynamic Stochastic Programming techniques to optimize the operation of the multipurpose reservoir system of the São Francisco River Basin. For this, the reservoir system was modeled by the Três Marias, Sobradinho and Itaparica mills. The developed model has three state variables and nine decision variables. The objective function was constructed using the weighting method, taking into account energy production and irrigation demands. The results showed better efficiency of Dual Dynamic Stochastic Programming that managed to optimize the complete hydrosystem for the various scenarios with low computational cost. The optimization used code developed in the Julia Language through the StochDynamicProgramming package. The language presented high level and high performance.