A Bayesian definition of 'most probable' parameters

Paolo Gardoni, Giovanna Biscontin

Research output: Contribution to journalReview article

Abstract

Since guidelines for choosing 'most probable' parameters in ground engineering design codes are vague, concerns are raised regarding their definition, as well as the associated uncertainties. This paper introduces Bayesian inference for a new rigorous approach to obtaining the estimates of the most probable parameters based on observations collected during construction. Following the review of optimisation-based methods that can be used in back-analysis, such as gradient descent and neural networks, a probabilistic model is developed using Clough and O'Rourke's method for retaining wall design. Sequential Bayesian inference is applied to a staged excavation project to examine the applicability of the proposed approach and illustrate the process of back-analysis.

Original languageEnglish (US)
Pages (from-to)130-142
Number of pages13
JournalGeotechnical Research
Volume5
Issue number3
DOIs
StatePublished - Sep 13 2018

Fingerprint

back analysis
Retaining walls
retaining wall
Excavation
excavation
Neural networks
engineering
method
parameter
code
project
Uncertainty
Statistical Models

ASJC Scopus subject areas

  • Geotechnical Engineering and Engineering Geology

Cite this

A Bayesian definition of 'most probable' parameters. / Gardoni, Paolo; Biscontin, Giovanna.

In: Geotechnical Research, Vol. 5, No. 3, 13.09.2018, p. 130-142.

Research output: Contribution to journalReview article

Gardoni, Paolo ; Biscontin, Giovanna. / A Bayesian definition of 'most probable' parameters. In: Geotechnical Research. 2018 ; Vol. 5, No. 3. pp. 130-142.
@article{b1d2f5ebd2aa446cb09bdea43559c419,
title = "A Bayesian definition of 'most probable' parameters",
abstract = "Since guidelines for choosing 'most probable' parameters in ground engineering design codes are vague, concerns are raised regarding their definition, as well as the associated uncertainties. This paper introduces Bayesian inference for a new rigorous approach to obtaining the estimates of the most probable parameters based on observations collected during construction. Following the review of optimisation-based methods that can be used in back-analysis, such as gradient descent and neural networks, a probabilistic model is developed using Clough and O'Rourke's method for retaining wall design. Sequential Bayesian inference is applied to a staged excavation project to examine the applicability of the proposed approach and illustrate the process of back-analysis.",
author = "Paolo Gardoni and Giovanna Biscontin",
year = "2018",
month = "9",
day = "13",
doi = "10.1680/jgere.18.00027",
language = "English (US)",
volume = "5",
pages = "130--142",
journal = "Geotechnical Research",
issn = "2052-6156",
publisher = "ICE Publishing Ltd.",
number = "3",

}

TY - JOUR

T1 - A Bayesian definition of 'most probable' parameters

AU - Gardoni, Paolo

AU - Biscontin, Giovanna

PY - 2018/9/13

Y1 - 2018/9/13

N2 - Since guidelines for choosing 'most probable' parameters in ground engineering design codes are vague, concerns are raised regarding their definition, as well as the associated uncertainties. This paper introduces Bayesian inference for a new rigorous approach to obtaining the estimates of the most probable parameters based on observations collected during construction. Following the review of optimisation-based methods that can be used in back-analysis, such as gradient descent and neural networks, a probabilistic model is developed using Clough and O'Rourke's method for retaining wall design. Sequential Bayesian inference is applied to a staged excavation project to examine the applicability of the proposed approach and illustrate the process of back-analysis.

AB - Since guidelines for choosing 'most probable' parameters in ground engineering design codes are vague, concerns are raised regarding their definition, as well as the associated uncertainties. This paper introduces Bayesian inference for a new rigorous approach to obtaining the estimates of the most probable parameters based on observations collected during construction. Following the review of optimisation-based methods that can be used in back-analysis, such as gradient descent and neural networks, a probabilistic model is developed using Clough and O'Rourke's method for retaining wall design. Sequential Bayesian inference is applied to a staged excavation project to examine the applicability of the proposed approach and illustrate the process of back-analysis.

UR - http://www.scopus.com/inward/record.url?scp=85053635374&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053635374&partnerID=8YFLogxK

U2 - 10.1680/jgere.18.00027

DO - 10.1680/jgere.18.00027

M3 - Review article

AN - SCOPUS:85053635374

VL - 5

SP - 130

EP - 142

JO - Geotechnical Research

JF - Geotechnical Research

SN - 2052-6156

IS - 3

ER -