Global sensitivity analysis for high-dimensional problems

How to objectively group factors and measure robustness and convergence while reducing computational cost

Razi Sheikholeslami, Saman Razavi, Hoshin Vijai Gupta, William Becker, Amin Haghnegahdar

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Dynamical earth and environmental systems models are typically computationally intensive and highly parameterized with many uncertain parameters. Together, these characteristics severely limit the applicability of Global Sensitivity Analysis (GSA) to high-dimensional models because very large numbers of model runs are typically required to achieve convergence and provide a robust assessment. Paradoxically, only 30 percent of GSA applications in the environmental modelling literature have investigated models with more than 20 parameters, suggesting that GSA is under-utilized on problems for which it should prove most useful. We develop a novel grouping strategy, based on bootstrap-based clustering, that enables efficient application of GSA to high-dimensional models. We also provide a new measure of robustness that assesses GSA stability and convergence. For two models, having 50 and 111 parameters, we show that grouping-enabled GSA provides results that are highly robust to sampling variability, while converging with a much smaller number of model runs.

Original languageEnglish (US)
Pages (from-to)282-299
Number of pages18
JournalEnvironmental Modelling and Software
Volume111
DOIs
StatePublished - Jan 1 2019

Fingerprint

Sensitivity analysis
sensitivity analysis
cost
Costs
environmental modeling
Earth (planet)
Sampling
sampling
parameter

Keywords

  • Computationally intensive simulations
  • Convergence
  • Curse of dimensionality
  • Factor grouping
  • Global sensitivity analysis
  • Robustness

ASJC Scopus subject areas

  • Software
  • Environmental Engineering
  • Ecological Modeling

Cite this

Global sensitivity analysis for high-dimensional problems : How to objectively group factors and measure robustness and convergence while reducing computational cost. / Sheikholeslami, Razi; Razavi, Saman; Gupta, Hoshin Vijai; Becker, William; Haghnegahdar, Amin.

In: Environmental Modelling and Software, Vol. 111, 01.01.2019, p. 282-299.

Research output: Contribution to journalArticle

@article{d7dc0dd94cd94f39bdefc4e7500fb64b,
title = "Global sensitivity analysis for high-dimensional problems: How to objectively group factors and measure robustness and convergence while reducing computational cost",
abstract = "Dynamical earth and environmental systems models are typically computationally intensive and highly parameterized with many uncertain parameters. Together, these characteristics severely limit the applicability of Global Sensitivity Analysis (GSA) to high-dimensional models because very large numbers of model runs are typically required to achieve convergence and provide a robust assessment. Paradoxically, only 30 percent of GSA applications in the environmental modelling literature have investigated models with more than 20 parameters, suggesting that GSA is under-utilized on problems for which it should prove most useful. We develop a novel grouping strategy, based on bootstrap-based clustering, that enables efficient application of GSA to high-dimensional models. We also provide a new measure of robustness that assesses GSA stability and convergence. For two models, having 50 and 111 parameters, we show that grouping-enabled GSA provides results that are highly robust to sampling variability, while converging with a much smaller number of model runs.",
keywords = "Computationally intensive simulations, Convergence, Curse of dimensionality, Factor grouping, Global sensitivity analysis, Robustness",
author = "Razi Sheikholeslami and Saman Razavi and Gupta, {Hoshin Vijai} and William Becker and Amin Haghnegahdar",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.envsoft.2018.09.002",
language = "English (US)",
volume = "111",
pages = "282--299",
journal = "Environmental Modelling and Software",
issn = "1364-8152",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Global sensitivity analysis for high-dimensional problems

T2 - How to objectively group factors and measure robustness and convergence while reducing computational cost

AU - Sheikholeslami, Razi

AU - Razavi, Saman

AU - Gupta, Hoshin Vijai

AU - Becker, William

AU - Haghnegahdar, Amin

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Dynamical earth and environmental systems models are typically computationally intensive and highly parameterized with many uncertain parameters. Together, these characteristics severely limit the applicability of Global Sensitivity Analysis (GSA) to high-dimensional models because very large numbers of model runs are typically required to achieve convergence and provide a robust assessment. Paradoxically, only 30 percent of GSA applications in the environmental modelling literature have investigated models with more than 20 parameters, suggesting that GSA is under-utilized on problems for which it should prove most useful. We develop a novel grouping strategy, based on bootstrap-based clustering, that enables efficient application of GSA to high-dimensional models. We also provide a new measure of robustness that assesses GSA stability and convergence. For two models, having 50 and 111 parameters, we show that grouping-enabled GSA provides results that are highly robust to sampling variability, while converging with a much smaller number of model runs.

AB - Dynamical earth and environmental systems models are typically computationally intensive and highly parameterized with many uncertain parameters. Together, these characteristics severely limit the applicability of Global Sensitivity Analysis (GSA) to high-dimensional models because very large numbers of model runs are typically required to achieve convergence and provide a robust assessment. Paradoxically, only 30 percent of GSA applications in the environmental modelling literature have investigated models with more than 20 parameters, suggesting that GSA is under-utilized on problems for which it should prove most useful. We develop a novel grouping strategy, based on bootstrap-based clustering, that enables efficient application of GSA to high-dimensional models. We also provide a new measure of robustness that assesses GSA stability and convergence. For two models, having 50 and 111 parameters, we show that grouping-enabled GSA provides results that are highly robust to sampling variability, while converging with a much smaller number of model runs.

KW - Computationally intensive simulations

KW - Convergence

KW - Curse of dimensionality

KW - Factor grouping

KW - Global sensitivity analysis

KW - Robustness

UR - http://www.scopus.com/inward/record.url?scp=85056209334&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056209334&partnerID=8YFLogxK

U2 - 10.1016/j.envsoft.2018.09.002

DO - 10.1016/j.envsoft.2018.09.002

M3 - Article

VL - 111

SP - 282

EP - 299

JO - Environmental Modelling and Software

JF - Environmental Modelling and Software

SN - 1364-8152

ER -