A new approach for measuring uncertainty importance and distributional sensitivity in probabilistic safety assessment

Park, Chang K.; Ahn, Kwang-Il

doi:10.1016/0951-8320(94)90119-8

Cited by 79 publications

(29 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…All three cases reflect the change in the state of knowledge of the analyst about the inputs. 25) Iði:oÞ represents the expected information for discrimination between f o ðyÞ and f i ðyÞ. The advantage of Iði:oÞ is that the entire output distribution is considered instead of a simple statistics of the output.…”

Section: Local Vs Global Importance Measures For Sensitivity Analmentioning

confidence: 99%

“…However, since the conditional distribution f i ðyÞ is dependent on the hypothesized change of an input, the result of Iði:oÞ is sensitive to how the distribution of the input is changed. In addition, since f o ðyÞ and f i ðyÞ are estimated based on the maximum entropy principle in Park and Ahn, 25) the use of data fitting procedures to derive f i ðyÞ and f o ðyÞ might lead to additional uncertainty.…”

Section: Local Vs Global Importance Measures For Sensitivity Analmentioning

confidence: 99%

“…From the definition of S ðCDFÞ i , S i , S Ti , and i , it is known that all the four importance measures are calculated without any assumption of the change in the input distributions. As to the measure Iði:oÞ, 25) since it is calculated with the presumption of the change in the distributions of the input parameters, the ranking of model inputs based on Iði:oÞ is sensitive to how the input distributions are changed. Hence, it is not meaningful to compare S ðCDFÞ i with Iði:oÞ directly.…”

Section: A Mathematical Modelmentioning

confidence: 99%

See 2 more Smart Citations

A New Importance Measure for Sensitivity Analysis

Homma

2010

Journal of Nuclear Science and Technology

View full text Add to dashboard Cite

Uncertainty is an integral part of risk assessment of complex engineering systems, such as nuclear power plants and space crafts. The aim of sensitivity analysis is to identify the contribution of the uncertainty in model inputs to the uncertainty in the model output. In this study, a new importance measure that characterizes the influence of the entire input distribution on the entire output distribution was proposed. It represents the expected deviation of the cumulative distribution function (CDF) of the model output that would be obtained when one input parameter of interest were known. The applicability of this importance measure was tested with two models, a nonlinear nonmonotonic mathematical model and a risk model. In addition, a comparison of this new importance measure with several other importance measures was carried out and the differences between these measures were explained.

show abstract

Section: Local Vs Global Importance Measures For Sensitivity Analmentioning

confidence: 99%

Section: Local Vs Global Importance Measures For Sensitivity Analmentioning

confidence: 99%

Section: A Mathematical Modelmentioning

confidence: 99%

See 1 more Smart Citation

A New Importance Measure for Sensitivity Analysis

Homma

2010

Journal of Nuclear Science and Technology

View full text Add to dashboard Cite

show abstract

“…27) Instead of using variance as the measure of uncertainty, Park and Ahn argued that uncertainty should be represented by a probability distribution, and the importance of an input parameter should reflect the degree of change in the whole output distribution. 25) They proposed to adopt the Kullback-Leibler discriminator 28) to measure the difference of one probability density function (PDF) f i ðyÞ from another f o ðyÞ:…”

Section: Local Vs Global Importance Measures For Sensitivity Analmentioning

confidence: 99%

“…Section II provides a description of several importance measures for sensitivity analysis, including the popularly used variance-based measures 7,10,[18][19][20][21][22][23] and moment-independent measures. 24,25) The strengths and limitations of these measures are pointed out. In Sec.…”

Section: Introductionmentioning

confidence: 99%

A New Importance Measure for Sensitivity Analysis

LIU

Homma

2010

J. Nucl. Sci. Technol.

View full text Add to dashboard Cite

show abstract

Distributed Evaluation of Local Sensitivity Analysis (DELSA), with application to hydrologic models

Rakovec¹,

et al. 2014

View full text Add to dashboard Cite

[1] This paper presents a hybrid local-global sensitivity analysis method termed the Distributed Evaluation of Local Sensitivity Analysis (DELSA), which is used here to identify important and unimportant parameters and evaluate how model parameter importance changes as parameter values change. DELSA uses derivative-based ''local'' methods to obtain the distribution of parameter sensitivity across the parameter space, which promotes consideration of sensitivity analysis results in the context of simulated dynamics. This work presents DELSA, discusses how it relates to existing methods, and uses two hydrologic test cases to compare its performance with the popular global, variance-based Sobol' method. The first test case is a simple nonlinear reservoir model with two parameters. The second test case involves five alternative ''bucket-style'' hydrologic models with up to 14 parameters applied to a medium-sized catchment (200 km 2 ) in the Belgian Ardennes. Results show that in both examples, Sobol' and DELSA identify similar important and unimportant parameters, with DELSA enabling more detailed insight at much lower computational cost. For example, in the real-world problem the time delay in runoff is the most important parameter in all models, but DELSA shows that for about 20% of parameter sets it is not important at all and alternative mechanisms and parameters dominate. Moreover, the time delay was identified as important in regions producing poor model fits, whereas other parameters were identified as more important in regions of the parameter space producing better model fits. The ability to understand how parameter importance varies through parameter space is critical to inform decisions about, for example, additional data collection and model development. The ability to perform such analyses with modest computational requirements provides exciting opportunities to evaluate complicated models as well as many alternative models.

show abstract

A new approach for measuring uncertainty importance and distributional sensitivity in probabilistic safety assessment

Cited by 79 publications

References 8 publications

A New Importance Measure for Sensitivity Analysis

A New Importance Measure for Sensitivity Analysis

A New Importance Measure for Sensitivity Analysis

Distributed Evaluation of Local Sensitivity Analysis (DELSA), with application to hydrologic models

Contact Info

Product

Resources

About