On the solution variability reduction of Stochastic Dual Dynamic Programming applied to energy planning

详细信息    查看全文

• 作者：Murilo Pereira Soares^a ; ^{murilopsoares@gmail.com} ; Alexandre Street^a ; ^{street@ele.puc-rio.br} ; Davi Michel Valladã ; o^b ; ^{davimv@puc-rio.br}
• 关键词：Stochastic programming ; Stochastic Dual Dynamic Programming ; Risk aversion ; OR in energy
• 刊名：European Journal of Operational Research
• 出版年：2017
• 出版时间：16 April 2017
• 年：2017
• 卷：258
• 期：2
• 页码：743-760
• 全文大小：4310 K
• 卷排序：258

文摘

SDDP is largely applied in hydrothermal operation planning, with its results impacting also market operations in Brazil. This work aims at regularizing SDDP solution variability when applied to hydrothermal energy planning problems. We propose two modeling changes to the hydrothermal problem to stabilize both primal and dual solutions. We show that it is possible to obtain solutions with similar characteristics as with traditional modeling, but with less variability.