On Regularization Schemes for Data-Driven Optimization

Wei Ni; Zhong Ping Jiang

doi:10.1109/CCDC.2019.8832868

On Regularization Schemes for Data-Driven Optimization

Wei Ni, Zhong Ping Jiang

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

This paper presents several regularization schemes for data-driven optimization problems. Data-driven optimization is a challenging task because the model or the distribution describing the inherent stochastic disturbance is unknown but instead only a set of data sampled from it can be available. Most existing literature tackles the data-driven optimization problem by using the approach of distributionally robust optimization which is however an infinite-dimensional optimization problem and thus is required to be transformed into a tractable finite-dimensional optimization problem. As a different line, the regularization schemes presented in this paper contribute to solving the data-driven minimization problem by approximating the unknown expectation with some data-dependent surrogates in a high confidence and by minimizing the surrogates in a tractable way. The tradeoff between the confidence level and the optimization error is analyzed.

Original language	English (US)
Title of host publication	Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	3016-3023
Number of pages	8
ISBN (Electronic)	9781728101057
DOIs	https://doi.org/10.1109/CCDC.2019.8832868
State	Published - Jun 2019
Event	31st Chinese Control and Decision Conference, CCDC 2019 - Nanchang, China Duration: Jun 3 2019 → Jun 5 2019

Publication series

Name	Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019

Conference

Conference	31st Chinese Control and Decision Conference, CCDC 2019
Country/Territory	China
City	Nanchang
Period	6/3/19 → 6/5/19

Keywords

Cressie-Read Likelihood
Data-driven optimization
chi-distribution
distributionally robust optimization
empirical likelihood
phi-divergence

ASJC Scopus subject areas

General Decision Sciences
Control and Systems Engineering
Control and Optimization

Access to Document

10.1109/CCDC.2019.8832868

Cite this

Ni, W., & Jiang, Z. P. (2019). On Regularization Schemes for Data-Driven Optimization. In Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019 (pp. 3016-3023). Article 8832868 (Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CCDC.2019.8832868

On Regularization Schemes for Data-Driven Optimization. / Ni, Wei; Jiang, Zhong Ping.
Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 3016-3023 8832868 (Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Ni, W & Jiang, ZP 2019, On Regularization Schemes for Data-Driven Optimization. in Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019., 8832868, Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019, Institute of Electrical and Electronics Engineers Inc., pp. 3016-3023, 31st Chinese Control and Decision Conference, CCDC 2019, Nanchang, China, 6/3/19. https://doi.org/10.1109/CCDC.2019.8832868

@inproceedings{51dd468c519c40ceb662b5c3a2a83766,

title = "On Regularization Schemes for Data-Driven Optimization",

abstract = "This paper presents several regularization schemes for data-driven optimization problems. Data-driven optimization is a challenging task because the model or the distribution describing the inherent stochastic disturbance is unknown but instead only a set of data sampled from it can be available. Most existing literature tackles the data-driven optimization problem by using the approach of distributionally robust optimization which is however an infinite-dimensional optimization problem and thus is required to be transformed into a tractable finite-dimensional optimization problem. As a different line, the regularization schemes presented in this paper contribute to solving the data-driven minimization problem by approximating the unknown expectation with some data-dependent surrogates in a high confidence and by minimizing the surrogates in a tractable way. The tradeoff between the confidence level and the optimization error is analyzed.",

keywords = "Cressie-Read Likelihood, Data-driven optimization, chi-distribution, distributionally robust optimization, empirical likelihood, phi-divergence",

author = "Wei Ni and Jiang, {Zhong Ping}",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 31st Chinese Control and Decision Conference, CCDC 2019 ; Conference date: 03-06-2019 Through 05-06-2019",

year = "2019",

month = jun,

doi = "10.1109/CCDC.2019.8832868",

language = "English (US)",

series = "Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "3016--3023",

booktitle = "Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019",

}

TY - GEN

T1 - On Regularization Schemes for Data-Driven Optimization

AU - Ni, Wei

AU - Jiang, Zhong Ping

PY - 2019/6

Y1 - 2019/6

N2 - This paper presents several regularization schemes for data-driven optimization problems. Data-driven optimization is a challenging task because the model or the distribution describing the inherent stochastic disturbance is unknown but instead only a set of data sampled from it can be available. Most existing literature tackles the data-driven optimization problem by using the approach of distributionally robust optimization which is however an infinite-dimensional optimization problem and thus is required to be transformed into a tractable finite-dimensional optimization problem. As a different line, the regularization schemes presented in this paper contribute to solving the data-driven minimization problem by approximating the unknown expectation with some data-dependent surrogates in a high confidence and by minimizing the surrogates in a tractable way. The tradeoff between the confidence level and the optimization error is analyzed.

AB - This paper presents several regularization schemes for data-driven optimization problems. Data-driven optimization is a challenging task because the model or the distribution describing the inherent stochastic disturbance is unknown but instead only a set of data sampled from it can be available. Most existing literature tackles the data-driven optimization problem by using the approach of distributionally robust optimization which is however an infinite-dimensional optimization problem and thus is required to be transformed into a tractable finite-dimensional optimization problem. As a different line, the regularization schemes presented in this paper contribute to solving the data-driven minimization problem by approximating the unknown expectation with some data-dependent surrogates in a high confidence and by minimizing the surrogates in a tractable way. The tradeoff between the confidence level and the optimization error is analyzed.

KW - Cressie-Read Likelihood

KW - Data-driven optimization

KW - chi-distribution

KW - distributionally robust optimization

KW - empirical likelihood

KW - phi-divergence

UR - http://www.scopus.com/inward/record.url?scp=85073096948&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073096948&partnerID=8YFLogxK

U2 - 10.1109/CCDC.2019.8832868

DO - 10.1109/CCDC.2019.8832868

M3 - Conference contribution

T3 - Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019

SP - 3016

EP - 3023

BT - Proceedings of the 31st Chinese Control and Decision Conference, CCDC 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 31st Chinese Control and Decision Conference, CCDC 2019

Y2 - 3 June 2019 through 5 June 2019

ER -

On Regularization Schemes for Data-Driven Optimization

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this