Improving the performance of Stochastic Dual Dynamic Programming

作者：

Highlights：

• Presents two tree traversing strategies for SDDP.

• Presents three cut selection algorithms to improve SDDP performance.

• Considers the whole Brazilian Power System in our computational results.

• Significant reduction in computational time without compromising the policy.

摘要

•Presents two tree traversing strategies for SDDP.•Presents three cut selection algorithms to improve SDDP performance.•Considers the whole Brazilian Power System in our computational results.•Significant reduction in computational time without compromising the policy.

论文关键词：Stochastic programming,Stochastic Dual Dynamic Programming,Cut selection,Hydrothermal scheduling

论文评审过程：Received 7 January 2013, Revised 8 January 2015, Available online 18 May 2015, Version of Record 9 June 2015.

论文官网地址：https://doi.org/10.1016/j.cam.2015.04.048