Robust propagation of probability boxes by interval predictor models

Jonathan Sadeghi, Marco de Angelis, Edoardo Patelli

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

This paper proposes numerical strategies to robustly and efficiently propagate probability boxes through expensive black box models. An interval is obtained for the system failure probability, with a confidence level. The three proposed algorithms are sampling based, and so can be easily parallelised, and make no assumptions about the functional form of the model. In the first two algorithms, the performance function is modelled as a function with unknown noise structure in the aleatory space and supplemented by a modified performance function. In the third algorithm, an Interval Predictor Model is constructed and a re-weighting strategy used to find bounds on the probability of failure. Numerical examples are presented to show the applicability of the approach. The proposed method is flexible and can account for epistemic uncertainty contained inside the limit state function. This is a feature which, to the best of the authors’ knowledge, no existing methods of this type can deal with.

Original languageEnglish
Article number101889
Number of pages10
JournalStructural Safety
Volume82
Early online date13 Sep 2019
DOIs
Publication statusE-pub ahead of print - 13 Sep 2019

Fingerprint

Sampling
Uncertainty

Keywords

  • imprecise probability
  • interval predictor models
  • probability boxes
  • reliability analysis
  • scenario optimisation

Cite this

@article{e6c03d1d00de4aa3b746fa617409b8a4,
title = "Robust propagation of probability boxes by interval predictor models",
abstract = "This paper proposes numerical strategies to robustly and efficiently propagate probability boxes through expensive black box models. An interval is obtained for the system failure probability, with a confidence level. The three proposed algorithms are sampling based, and so can be easily parallelised, and make no assumptions about the functional form of the model. In the first two algorithms, the performance function is modelled as a function with unknown noise structure in the aleatory space and supplemented by a modified performance function. In the third algorithm, an Interval Predictor Model is constructed and a re-weighting strategy used to find bounds on the probability of failure. Numerical examples are presented to show the applicability of the approach. The proposed method is flexible and can account for epistemic uncertainty contained inside the limit state function. This is a feature which, to the best of the authors’ knowledge, no existing methods of this type can deal with.",
keywords = "imprecise probability, interval predictor models, probability boxes, reliability analysis, scenario optimisation",
author = "Jonathan Sadeghi and {de Angelis}, Marco and Edoardo Patelli",
year = "2019",
month = "9",
day = "13",
doi = "10.1016/j.strusafe.2019.101889",
language = "English",
volume = "82",
journal = "Structural Safety",
issn = "0167-4730",
publisher = "Elsevier",

}

Robust propagation of probability boxes by interval predictor models. / Sadeghi, Jonathan; de Angelis, Marco; Patelli, Edoardo.

In: Structural Safety, Vol. 82, 101889, 31.01.2020.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Robust propagation of probability boxes by interval predictor models

AU - Sadeghi, Jonathan

AU - de Angelis, Marco

AU - Patelli, Edoardo

PY - 2019/9/13

Y1 - 2019/9/13

N2 - This paper proposes numerical strategies to robustly and efficiently propagate probability boxes through expensive black box models. An interval is obtained for the system failure probability, with a confidence level. The three proposed algorithms are sampling based, and so can be easily parallelised, and make no assumptions about the functional form of the model. In the first two algorithms, the performance function is modelled as a function with unknown noise structure in the aleatory space and supplemented by a modified performance function. In the third algorithm, an Interval Predictor Model is constructed and a re-weighting strategy used to find bounds on the probability of failure. Numerical examples are presented to show the applicability of the approach. The proposed method is flexible and can account for epistemic uncertainty contained inside the limit state function. This is a feature which, to the best of the authors’ knowledge, no existing methods of this type can deal with.

AB - This paper proposes numerical strategies to robustly and efficiently propagate probability boxes through expensive black box models. An interval is obtained for the system failure probability, with a confidence level. The three proposed algorithms are sampling based, and so can be easily parallelised, and make no assumptions about the functional form of the model. In the first two algorithms, the performance function is modelled as a function with unknown noise structure in the aleatory space and supplemented by a modified performance function. In the third algorithm, an Interval Predictor Model is constructed and a re-weighting strategy used to find bounds on the probability of failure. Numerical examples are presented to show the applicability of the approach. The proposed method is flexible and can account for epistemic uncertainty contained inside the limit state function. This is a feature which, to the best of the authors’ knowledge, no existing methods of this type can deal with.

KW - imprecise probability

KW - interval predictor models

KW - probability boxes

KW - reliability analysis

KW - scenario optimisation

UR - http://www.scopus.com/inward/record.url?scp=85072208904&partnerID=8YFLogxK

UR - https://livrepository.liverpool.ac.uk/3052722/

U2 - 10.1016/j.strusafe.2019.101889

DO - 10.1016/j.strusafe.2019.101889

M3 - Article

VL - 82

JO - Structural Safety

JF - Structural Safety

SN - 0167-4730

M1 - 101889

ER -