Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty

A. Gray; A. Wimbush; M. Deangelis; P. O. Hristov; E. Miralles-Dolz; D. Calleja; R. Rocchetta

doi:10.3850/978-981-14-8593-0_5520-cd

Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty

A. Gray
, A. Wimbush
, M. Deangelis
, P. O. Hristov
, E. Miralles-Dolz
, D. Calleja
, R. Rocchetta

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

Abstract

Uncertainty quantification is a vital part of all engineering and scientific pursuits. Some of the current most challenging tasks in UQ involve accurately calibrating, propagating and performing optimisation under aleatory and epistemic uncertainty in high dimensional models with very few data; like the challenge proposed by Nasa Langley this year. In this paper we propose a solution which clearly separates aleatory from epistemic uncertainty. A multidimensional 2nd-order distribution was calibrated with Bayesian updating and used as an inner approximation to a p-box. A sliced normal distribution was fit to the posterior, and used to produce cheap samples while keeping the posterior dependence structure. The remaining tasks, such as sensitivity and reliability optimisation, are completed with probability bounds analysis. These tasks were repeated a number of times as designs were improved and more data gathered.

Original language	American English
Title of host publication	Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference
Editors	Piero Baraldi, Francesco Di Maio, Enrico Zio
Publisher	Research Publishing, Singapore
Pages	1111-1118
Number of pages	8
ISBN (Print)	9789811485930
DOIs	https://doi.org/10.3850/978-981-14-8593-0_5520-cd
State	Published - 2020
Externally published	Yes
Event	30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020 - Venice, Italy Duration: Nov 1 2020 → Nov 5 2020

Publication series

Name	Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference

Conference

Conference	30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020
Country/Territory	Italy
City	Venice
Period	11/1/20 → 11/5/20

ASJC Scopus subject areas

Safety, Risk, Reliability and Quality
Safety Research

Keywords

2nd-order distribution
Bayesian calibration
Epistemic uncertainty
Optimization under uncertainty
Probability bounds analysis
Uncertainty propagation
Uncertainty reduction

Access to Document

10.3850/978-981-14-8593-0_5520-cd

Cite this

Gray, A., Wimbush, A., Deangelis, M., Hristov, P. O., Miralles-Dolz, E., Calleja, D., & Rocchetta, R. (2020). Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty. In P. Baraldi, F. Di Maio, & E. Zio (Eds.), Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference (pp. 1111-1118). (Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference). Research Publishing, Singapore. https://doi.org/10.3850/978-981-14-8593-0_5520-cd

Gray, A. ; Wimbush, A. ; Deangelis, M. et al. / Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty. Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference. editor / Piero Baraldi ; Francesco Di Maio ; Enrico Zio. Research Publishing, Singapore, 2020. pp. 1111-1118 (Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference).

@inproceedings{13b0eeffc77d4cdc837dbe466732b286,

title = "Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty",

abstract = "Uncertainty quantification is a vital part of all engineering and scientific pursuits. Some of the current most challenging tasks in UQ involve accurately calibrating, propagating and performing optimisation under aleatory and epistemic uncertainty in high dimensional models with very few data; like the challenge proposed by Nasa Langley this year. In this paper we propose a solution which clearly separates aleatory from epistemic uncertainty. A multidimensional 2nd-order distribution was calibrated with Bayesian updating and used as an inner approximation to a p-box. A sliced normal distribution was fit to the posterior, and used to produce cheap samples while keeping the posterior dependence structure. The remaining tasks, such as sensitivity and reliability optimisation, are completed with probability bounds analysis. These tasks were repeated a number of times as designs were improved and more data gathered.",

keywords = "2nd-order distribution, Bayesian calibration, Epistemic uncertainty, Optimization under uncertainty, Probability bounds analysis, Uncertainty propagation, Uncertainty reduction",

author = "A. Gray and A. Wimbush and M. Deangelis and Hristov, \{P. O.\} and E. Miralles-Dolz and D. Calleja and R. Rocchetta",

note = "Publisher Copyright: {\textcopyright} ESREL2020-PSAM15 Organizers.Published by Research Publishing, Singapore.; 30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020 ; Conference date: 01-11-2020 Through 05-11-2020",

year = "2020",

doi = "10.3850/978-981-14-8593-0\_5520-cd",

language = "American English",

isbn = "9789811485930",

series = "Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference",

publisher = "Research Publishing, Singapore",

pages = "1111--1118",

editor = "Piero Baraldi and \{Di Maio\}, Francesco and Enrico Zio",

booktitle = "Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference",

}

Gray, A, Wimbush, A, Deangelis, M, Hristov, PO, Miralles-Dolz, E, Calleja, D & Rocchetta, R 2020, Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty. in P Baraldi, F Di Maio & E Zio (eds), Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference. Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference, Research Publishing, Singapore, pp. 1111-1118, 30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020, Venice, Italy, 11/1/20. https://doi.org/10.3850/978-981-14-8593-0_5520-cd

Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty. / Gray, A.; Wimbush, A.; Deangelis, M. et al.
Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference. ed. / Piero Baraldi; Francesco Di Maio; Enrico Zio. Research Publishing, Singapore, 2020. p. 1111-1118 (Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference).

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

TY - GEN

T1 - Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty

AU - Gray, A.

AU - Wimbush, A.

AU - Deangelis, M.

AU - Hristov, P. O.

AU - Miralles-Dolz, E.

AU - Calleja, D.

AU - Rocchetta, R.

N1 - Publisher Copyright: © ESREL2020-PSAM15 Organizers.Published by Research Publishing, Singapore.

PY - 2020

Y1 - 2020

N2 - Uncertainty quantification is a vital part of all engineering and scientific pursuits. Some of the current most challenging tasks in UQ involve accurately calibrating, propagating and performing optimisation under aleatory and epistemic uncertainty in high dimensional models with very few data; like the challenge proposed by Nasa Langley this year. In this paper we propose a solution which clearly separates aleatory from epistemic uncertainty. A multidimensional 2nd-order distribution was calibrated with Bayesian updating and used as an inner approximation to a p-box. A sliced normal distribution was fit to the posterior, and used to produce cheap samples while keeping the posterior dependence structure. The remaining tasks, such as sensitivity and reliability optimisation, are completed with probability bounds analysis. These tasks were repeated a number of times as designs were improved and more data gathered.

AB - Uncertainty quantification is a vital part of all engineering and scientific pursuits. Some of the current most challenging tasks in UQ involve accurately calibrating, propagating and performing optimisation under aleatory and epistemic uncertainty in high dimensional models with very few data; like the challenge proposed by Nasa Langley this year. In this paper we propose a solution which clearly separates aleatory from epistemic uncertainty. A multidimensional 2nd-order distribution was calibrated with Bayesian updating and used as an inner approximation to a p-box. A sliced normal distribution was fit to the posterior, and used to produce cheap samples while keeping the posterior dependence structure. The remaining tasks, such as sensitivity and reliability optimisation, are completed with probability bounds analysis. These tasks were repeated a number of times as designs were improved and more data gathered.

KW - 2nd-order distribution

KW - Bayesian calibration

KW - Epistemic uncertainty

KW - Optimization under uncertainty

KW - Probability bounds analysis

KW - Uncertainty propagation

KW - Uncertainty reduction

UR - https://www.scopus.com/pages/publications/85107279115

UR - https://www.scopus.com/pages/publications/85107279115#tab=citedBy

U2 - 10.3850/978-981-14-8593-0_5520-cd

DO - 10.3850/978-981-14-8593-0_5520-cd

M3 - Conference contribution

SN - 9789811485930

T3 - Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference

SP - 1111

EP - 1118

BT - Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference

A2 - Baraldi, Piero

A2 - Di Maio, Francesco

A2 - Zio, Enrico

PB - Research Publishing, Singapore

T2 - 30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020

Y2 - 1 November 2020 through 5 November 2020

ER -

Gray A, Wimbush A, Deangelis M, Hristov PO, Miralles-Dolz E, Calleja D et al. Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty. In Baraldi P, Di Maio F, Zio E, editors, Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference. Research Publishing, Singapore. 2020. p. 1111-1118. (Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Management Conference). doi: 10.3850/978-981-14-8593-0_5520-cd

Bayesian calibration and probability bounds analysis solution to the nasa 2020 uq challenge on optimization under uncertainty

Abstract

Publication series

Conference

ASJC Scopus subject areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this