A Study on Sampling Stochastic Dynamic Programming for Long-term Hydropower Operation

Long-term hydropower operation is a complex optimization problem, as the uncertainty of natural inflow should be considered. Sampling stochastic dynamic programming (SSDP) is a method that based on the classical stochastic dynamic programming (SDP), and has the potential to capture the consecutive i...

Full description

Saved in:
Bibliographic Details
Published in:2025 7th International Conference on Energy Systems and Electrical Power (ICESEP) pp. 622 - 626
Main Authors: Weng, Shuqin, Kang, Chuanxiong, Zhao, Zenghai, Zhu, Fangliang, Gao, Jie, Wang, Xu, Guo, Peng
Format: Conference Proceeding
Language:English
Published: IEEE 20.06.2025
Subjects:
Dynamic programming
Hydroelectric power generation
long-term hydropower operation
Optimization
sampling stochastic dynamic programming
Shape
Simulation
streamflow scenarios
transition probabilities
Uncertainty
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Long-term hydropower operation is a complex optimization problem, as the uncertainty of natural inflow should be considered. Sampling stochastic dynamic programming (SSDP) is a method that based on the classical stochastic dynamic programming (SDP), and has the potential to capture the consecutive information of the natural inflow series. This paper studies the SSDP for long-term hydropower operation, and presents a method to determine transition probabilities by taking into account the similarity of streamflow scenarios in both total volume and process shape, to which weights are assigned. The SSDP, with the transition probabilities determined by the proposed method, is applied in two case studies, which suggests that the SSDP, for all weights tested in determining the transition probabilities, performs better than the SDP. The paper also suggests how to select the weights to make the procedure perform even better.
DOI:10.1109/ICESEP66633.2025.11155439