Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

Lucas Merabet; Bernardo Freitas Paulo da Costa; Vincent Leclere

doi:10.1007/s10287-024-00524-z

Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

Lucas Merabet
, Bernardo Freitas Paulo da Costa
, Vincent Leclere

Research output: Contribution to journal › Article › peer-review

Abstract

Risk-averse multistage problems and their applications are gaining interest in various fields of applications. Under convexity assumptions, the resolution of these problems can be done with trajectory following dynamic programming algorithms like Stochastic Dual Dynamic Programming (SDDP) to access a deterministic lower bound, and dual SDDP for deterministic upper bounds. In this paper, we leverage the dual SDDP algorithm to compute a policy with guaranteed risk-adjusted performance for multistage stochastic linear problems.

Original language	English
Article number	43
Journal	Computational Management Science
Volume	21
Issue number	2
DOIs	https://doi.org/10.1007/s10287-024-00524-z
Publication status	Published - 1 Dec 2024

Keywords

Duality
Inner approximation
Primal-dual methods
SDDP

Access to Document

10.1007/s10287-024-00524-z

Cite this

@article{78fe6888c3c54eb5bcdb03119c8c5081,

title = "Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems",

abstract = "Risk-averse multistage problems and their applications are gaining interest in various fields of applications. Under convexity assumptions, the resolution of these problems can be done with trajectory following dynamic programming algorithms like Stochastic Dual Dynamic Programming (SDDP) to access a deterministic lower bound, and dual SDDP for deterministic upper bounds. In this paper, we leverage the dual SDDP algorithm to compute a policy with guaranteed risk-adjusted performance for multistage stochastic linear problems.",

keywords = "Duality, Inner approximation, Primal-dual methods, SDDP",

author = "Lucas Merabet and \{da Costa\}, \{Bernardo Freitas Paulo\} and Vincent Leclere",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.",

year = "2024",

month = dec,

day = "1",

doi = "10.1007/s10287-024-00524-z",

language = "English",

volume = "21",

journal = "Computational Management Science",

issn = "1619-697X",

number = "2",

}

TY - JOUR

T1 - Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

AU - Merabet, Lucas

AU - da Costa, Bernardo Freitas Paulo

AU - Leclere, Vincent

N1 - Publisher Copyright: © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

PY - 2024/12/1

Y1 - 2024/12/1

N2 - Risk-averse multistage problems and their applications are gaining interest in various fields of applications. Under convexity assumptions, the resolution of these problems can be done with trajectory following dynamic programming algorithms like Stochastic Dual Dynamic Programming (SDDP) to access a deterministic lower bound, and dual SDDP for deterministic upper bounds. In this paper, we leverage the dual SDDP algorithm to compute a policy with guaranteed risk-adjusted performance for multistage stochastic linear problems.

AB - Risk-averse multistage problems and their applications are gaining interest in various fields of applications. Under convexity assumptions, the resolution of these problems can be done with trajectory following dynamic programming algorithms like Stochastic Dual Dynamic Programming (SDDP) to access a deterministic lower bound, and dual SDDP for deterministic upper bounds. In this paper, we leverage the dual SDDP algorithm to compute a policy with guaranteed risk-adjusted performance for multistage stochastic linear problems.

KW - Duality

KW - Inner approximation

KW - Primal-dual methods

KW - SDDP

UR - https://www.scopus.com/pages/publications/85204895088

U2 - 10.1007/s10287-024-00524-z

DO - 10.1007/s10287-024-00524-z

M3 - Article

AN - SCOPUS:85204895088

SN - 1619-697X

VL - 21

JO - Computational Management Science

JF - Computational Management Science

IS - 2

M1 - 43

ER -

Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this