Controlled approximation of the value function in stochastic dynamic programming for multi-reservoir systems

Authors: Zéphyr, LucknyLang, PascalLamond, Bernard
Abstract: We present a new approach for adaptive approximation of the value function in stochastic dynamic programming. Under convexity assumptions, our method is based on a simplicial partition of the state space. Bounds on the value function provide guidance as to where refinement should be done, if at all. Thus, the method allows for a trade-off between solution time and accuracy. The proposed scheme is experimented in the particular context of hydroelectric production across multiple reservoirs.
Document Type: Article de recherche
Issue Date: 1 October 2015
Open Access Date: 1 October 2016
Document version: AM
Permalink: http://hdl.handle.net/20.500.11794/10448
This document was published in: Computational Management Science, Vol. 12 (4), 539–557 (2015)
https://doi.org/10.1007/s10287-015-0242-1
Springer
Alternative version: 10.1007/s10287-015-0242-1
Collection:Articles publiés dans des revues avec comité de lecture

Files in this item:
SizeFormat 
CMSC-D-14-00044-xxx.pdf441.59 kBAdobe PDFView/Open
All documents in CorpusUL are protected by Copyright Act of Canada.