Sensitivity analysis: A review of recent advances

Published on Feb 1, 2016in European Journal of Operational Research3.806
· DOI :10.1016/j.ejor.2015.06.032
Emanuele Borgonovo26
Estimated H-index: 26
(Bocconi University),
Elmar Plischke10
Estimated H-index: 10
(TUC: Clausthal University of Technology)
The solution of several operations research problems requires the creation of a quantitative model. Sensitivity analysis is a crucial step in the model building and result communication process. Through sensitivity analysis we gain essential insights on model behavior, on its structure and on its response to changes in the model inputs. Several interrogations are possible and several sensitivity analysis methods have been developed, giving rise to a vast and growing literature. We present an overview of available methods, structuring them into local and global methods. For local methods, we discuss Tornado diagrams, one way sensitivity functions, differentiation-based methods and scenario decomposition through finite change sensitivity indices, providing a unified view of the associated sensitivity measures. We then analyze global sensitivity methods, first discussing screening methods such as sequential bifurcation and the Morris method. We then address variance-based, moment-independent and value of information-based sensitivity methods. We discuss their formalization in a common rationale and present recent results that permit the estimation of global sensitivity measures by post-processing the sample generated by a traditional Monte Carlo simulation. We then investigate in detail the methodological issues concerning the crucial step of correctly interpreting the results of a sensitivity analysis. A classical example is worked out to illustrate some of the approaches.
  • References (192)
  • Citations (212)
📖 Papers frequently viewed together
421 Citations
1,873 Citations
933 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
ABSTRACTPhysical phenomena are commonly modelled by time consuming numerical simulators, function of many uncertain parameters whose influences can be measured via a global sensitivity analysis. The usual variance-based indices require too many simulations, especially as the inputs are numerous. To address this limitation, we consider recent advances in dependence measures, focusing on the distance correlation and the Hilbert–Schmidt independence criterion. We study and use these indices for a s...
15 CitationsSource
#1John D. C. Little (MIT: Massachusetts Institute of Technology)H-Index: 36
(This article originally appeared in Management Science, April 1970, Volume 16, Number 8, pp. B-466-B-485, published by The Institute of Management Sciences.) A manager tries to put together the various resources under his control into an activity that achieves his objectives. A model of his operation can assist him but probably will not unless it meets certain requirements. A model that is to be used by a manager should be simple, robust, easy to control, adaptive, as complete as possible, and ...
766 Citations
3,019 Citations
Global sensitivity analysis with variance-based measures suffers from several theoretical and practical limitations, since they focus only on the variance of the output and handle multivariate variables in a limited way. In this paper, we introduce a new class of sensitivity indices based on dependence measures which overcomes these insufficiencies. Our approach originates from the idea to compare the output distribution with its conditional counterpart when one of the input variables is fixed. ...
48 CitationsSource
#1Qingqing Zhai (Beihang University)H-Index: 14
#2Jun Yang (Beihang University)H-Index: 13
Last. Yu Zhao (Beihang University)H-Index: 10
view all 4 authors...
Importance measures have been widely studied and applied in reliability and safety engineering. This paper presents a general formulation of moment-independent importance measures and several commonly discussed importance measures are unified based on Minkowski distance (MD). Moment-independent importance measures can be categorized into three classes of MD importance measures, i.e. probability density function based MD importance measure, cumulative distribution function based MD importance mea...
16 CitationsSource
#1Leigang Zhang (NPU: Northwestern Polytechnical University)H-Index: 3
#2Zhenzhou Lu (NPU: Northwestern Polytechnical University)H-Index: 21
Last. Chongqing Fan (NPU: Northwestern Polytechnical University)H-Index: 1
view all 4 authors...
The moment-independent importance measure proposed by Borgonovo, which is defined as the average shift between the unconditional and conditional probability density functions (PDFs) of model output, is widely used to evaluate the influence of input uncertainty on the entire output distribution. And how to exactly and efficiently estimate the PDFs remains a crucial and challenging problem. In this paper, a novel PDF estimation based method is proposed to efficiently evaluate the moment-independen...
27 CitationsSource
#1Emanuele Borgonovo (Bocconi University)H-Index: 26
#2Stefano TarantolaH-Index: 45
Last. Max D. Morris (Iowa State University)H-Index: 20
view all 4 authors...
type="main" xml:id="rssb12052-abs-0001"> Monotonic transformations are widely employed in statistics and data analysis. In computer experiments they are often used to gain accuracy in the estimation of global sensitivity statistics. However, one faces the question of interpreting results that are obtained on the transformed data back on the original data. The situation is even more complex in computer experiments, because transformations alter the model input–output mapping and distort the estim...
45 CitationsSource
#1Qingqing Zhai (Beihang University)H-Index: 14
#2Jun Yang (Beihang University)H-Index: 13
Last. Yu Zhao (Beihang University)H-Index: 10
view all 3 authors...
Abstract Variance-based sensitivity analysis has been widely studied and asserted itself among practitioners. Monte Carlo simulation methods are well developed in the calculation of variance-based sensitivity indices but they do not make full use of each model run. Recently, several works mentioned a scatter-plot partitioning method to estimate the variance-based sensitivity indices from given data, where a single bunch of samples is sufficient to estimate all the sensitivity indices. This paper...
17 CitationsSource
#1Leopoldo Eduardo Cárdenas-Barrón (Tec: Monterrey Institute of Technology and Higher Education)H-Index: 37
#2Kun-Jen Chung (Shih Chien University)H-Index: 18
Last. Gerardo Treviño-Garza (IBS: International Business School, Germany)H-Index: 12
view all 3 authors...
Abstract This writing presents a brief introduction to the papers included in the special issue “Celebrating a century of the economic order quantity model in honor of Ford Whitman Harris” published by the International Journal of Production Economics. Forty-one papers covering an extensive scope of inventory management have been incorporated in this volume from contributing authors from 20 countries located in America, Asia, Europe and Africa. This special issue also provides a basis for new di...
107 CitationsSource
#1Daniel Straub (TUM: Technische Universität München)H-Index: 24
Abstract When designing monitoring systems and planning inspections, engineers must assess the benefits of the additional information that can be obtained and weigh them against the cost of these measures. The value of information (VoI) concept of the Bayesian statistical decision analysis provides a formal framework to quantify these benefits. This paper presents the determination of the VoI when information is collected to increase the reliability of engineering systems. It is demonstrated how...
40 CitationsSource
Cited By212
Abstract In this study fundamental models were used by considering heat and mass multiphase transport in a porous medium which was coupled with the intermittent microwave convective combination drying process. The models were validated by experimental data sourced from references, and independent analysis was conducted to select a suitable mesh and time step. Sensitivity analysis was performed using different empirical formula-type input parameters. The water activity and the thermal conductivit...
Abstract A key source of uncertainty in the environmental assessment of emerging technologies is the unpredictable manufacturing, use, and end-of-life pathways a technology can take as it progresses from lab to industrial scale. This uncertainty has sometimes been addressed in life cycle assessment (LCA) by performing scenario analysis. However, the scenario-based approach can be misleading if the probabilities of occurrence of each scenario are not incorporated. It also brings about a practical...
#1Max Ehre (TUM: Technische Universität München)H-Index: 1
#2Iason Papaioannou (TUM: Technische Universität München)H-Index: 10
Last. Daniel Straub (TUM: Technische Universität München)H-Index: 24
view all 3 authors...
Abstract Global sensitivity analysis is a central part of uncertainty quantification with engineering models. Variance-based sensitivity measures such as Sobol’ and total-effect indices are amongst the most popular and commonly used tools for global sensitivity analysis. Mutiple sampling-based estimators of these measures are available, but they often come at considerable computational cost due to the large number of required model evaluations. If the computational model is expensive to evaluate...
#1Ji-Su Kim (Yonsei University)
#2Se Yun Kim (Yonsei University)H-Index: 3
Last. Tongseok Han (Yonsei University)H-Index: 15
view all 3 authors...
Abstract The spatial distribution of microstructures in cement paste affects its mechanical and thermal properties. To evaluate the material properties of cement pastes statistically, sensitivity analysis using the first-order second-moment (FOSM) method is performed. Two random variables, namely porosity and pore continuity, are selected as input variables, and the stiffness and thermal conductivity of cement paste are selected as output variables. The statistical distributions of the input var...
#1Thomas Høgholm Jørgensen (UCPH: University of Copenhagen)H-Index: 3
Across many fields in economics, a common approach to estimation of economic models is to calibrate a sub-set of model parameters and keep them fixed when estimating the remaining parameters. Calibrated parameters likely affect conclusions based on the model but estimation time often makes a systematic investigation of the sensitivity to calibrated parameters infeasible. I propose a simple and computationally low-cost measure of the sensitivity of parameters and other objects of interest to the ...
#1John McFarland (SwRI: Southwest Research Institute)H-Index: 8
#2Erin C. DeCarlo (SwRI: Southwest Research Institute)
Abstract Probabilistic methods are used with modeling and simulation to predict variation in system performance and assess risk due to randomness in model inputs such as material properties, loads, and boundary conditions. It is common practice to assume that the input distributions are known. However, this discounts the epistemic uncertainty in the values of the distribution parameters, which can be attributed to the availability of limited data to define the input distributions. This paper pro...
#1Li Lei (NPU: Northwestern Polytechnical University)
#2Chang Qi (NPU: Northwestern Polytechnical University)
Last. Zhang Zheng (NPU: Northwestern Polytechnical University)
view all 5 authors...
The complex layout and harsh working environment often encountered in aeronautical hydraulic pipeline system (AHPS) result in the requirement of a large number of constraints that are mainly utiliz...
#1Yiqi Liu (SCUT: South China University of Technology)
#2Min Xie (CityU: City University of Hong Kong)
Abstract Soft-sensors usually assist in dealing with the unavailability of hardware sensors in process industries, thus allowing for less fault occurrence and better control performance. However, nonlinear, non-stationary, ill-data, auto-correlated and co-correlated behaviors in industrial data always make general data-driven methods inadequate, thus resorting to kernel-based methods provide a necessary alternative. This paper gives a systematic review of various state-of-the-art kernel-based me...
#1Carlo Alberto Magni (University of Modena and Reggio Emilia)H-Index: 15
#2Andrea Marchioni (University of Modena and Reggio Emilia)
Abstract In project appraisal under uncertainty, the economic reliability of a measure of financial efficiency such as a rate of return depends on its strong NPV-consistency, meaning that the performance metric (i) supplies the same recommendation in accept-reject decisions as the NPV, (ii) ranks competing projects in the same way as the NPV, (iii) has the same sensitivity to perturbations in the input data as the NPV. In real-life projects, financial efficiency is greatly affected by the manage...
#1Junfei QiaoH-Index: 1
#2Xin GuoH-Index: 1
Last. Wenjing LiH-Index: 1
view all 3 authors...