Abstract
The choice of an adequate mathematical model is a key step in solving problems in many different fields. When more than one model is available to represent a given phenomenon, a poor choice might result in loss of precision and efficiency. Well-known strategies for comparing mathematical models can be found in many previous works, but seldom regarding several models with uncertain parameters at once. In this work, we present a novel approach for measuring the similarity among any given number of mathematical models, so as to support decision making regarding model selection. The strategy consists in defining a new general model composed of all candidate models and a uniformly distributed random variable, whose sampling selects the candidate model employed to evaluate the response. Global Sensitivity Analysis (GSA) is then performed to measure the sensitivity of the response with respect to this random variable. The result indicates the level of discrepancy among the mathematical models in the stochastic context. We also demonstrate that the proposed approach is related to the Root Mean Square (RMS) error when only two models are compared. The main advantages of the proposed approach are: (i) the problem is cast in the sound framework of GSA, (ii) the approach also quantifies if the discrepancy among the mathematical models is significant in comparison to uncertainties/randomness of the parameters, an analysis that is not possible with RMS error alone. Numerical examples of different disciplines and degrees of complexity are presented, showing the kind of insight we can get from the proposed approach.
Similar content being viewed by others
Data availability
All data used in this paper were obtained numerically. The first two examples can be reproduced with the details described in the text. The reader can contact the corresponding author in order to obtain the structural model employed in the third example.
Notes
Here, we employ the term mathematical model to represent the mathematical/numerical relation between input \({\textbf{x}} \in \mathbb {R}^n\) and output \({\textbf{y}} \in \mathbb {R}\).
In this work, we do not consider Sobol’ first order index \(S_W = \mathbb {V}\left[ \mathbb {E}\left[ Y \vert W\right] \right] /\mathbb {V}\left[ Y\right] \) because the total index \(S_{TW}\) is more appropriate for factor fixing decision (Saltelli et al. 2007)
References
Adnan RM, Petroselli A, Heddam S, Santos CAG, Kisi O (2021) Short term rainfall-runoff modelling using several machine learning methods and a conceptual event-based model. Stochastic Environ Res Risk Assess 35:597–616. https://doi.org/10.1007/s00477-020-01910-0
Alberg H, Berglund D (2003) Comparison of plastic, viscoplastic, and creep models when modelling welding and stress relief heat treatment. Comput Methods Appl Mech Eng 192(49):5189–5208. https://doi.org/10.1016/j.cma.2003.07.010
Allen CB, Taylor NV, Fenwick CL, Gaitonde AL, Jones DP (2005) A comparison of full non-linear and reduced order aerodynamic models in control law design using a two-dimensional aerofoil model. Int J Nume Methods Eng 64(12):1628–1648. https://doi.org/10.1002/nme.1421
Antoniadis A, Lambert-Lacroix S, Poggi J-M (2021) Random forests for global sensitivity analysis: a selective review. Reliab Eng Syst Saf 206:107312. https://doi.org/10.1016/j.ress.2020.107312
Bauer J, Tyacke J (2022) Comparison of low reynolds number turbulence and conjugate heat transfer modelling for pin-fin roughness elements. Appl Math Model 103:696–713. https://doi.org/10.1016/j.apm.2021.10.044
Begnini R, Torii AJ, Kroetz HM (2022) Mechanical model comparison using sobol’ indices. In: Proceedings of CILAMCE 2022
Borgonovo E (2010) Sensitivity analysis with finite changes: An application to modified eoq models. Euro J Oper Res 200(1):127–138. https://doi.org/10.1016/j.ejor.2008.12.025
Borgonovo E, Plischke E (2016) Sensitivity analysis: a review of recent advances. Euro J Oper Res 248(3):869–887. https://doi.org/10.1016/j.ejor.2015.06.032
Cannavó F (2012) Sensitivity analysis for volcanic source modeling quality assessment and model selection. Comput Geosci 44:52–59. https://doi.org/10.1016/j.cageo.2012.03.008
Cazes F, Moës N (2015) Comparison of a phase-field model and of a thick level set model for brittle and quasi-brittle fracture. Int J Numer Methods Eng 103(2):114–143. https://doi.org/10.1002/nme.4886
de Assis LME, Banerjee M, Venturino E (2020) Comparison of hidden and explicit resources in ecoepidemic models of predator-prey type. Comput Appl Math 39:36. https://doi.org/10.1007/s40314-019-1015-1
den Boon S, Jit M, Brisson M, Medley G, Beutels P, White R, Flasche S, Hollingsworth TD, Garske T, Pitzer VE, Hoogendoorn M, Geffen O, Clark A, Kim J, Hutubessy R (2019) Guidelines for multi-model comparisons of the impact of infectious disease interventions. BMC Medicine 17(163). https://doi.org/10.1186/s12916-019-1403-9
Ehre M, Papaioannou I, Straub D (2020) Global sensitivity analysis in high dimensions with pls-pce. Reliab Eng Syst Saf 198:106861. https://doi.org/10.1016/j.ress.2020.106861
Fort J-C, Klein T, Rachdi N (2016) New sensitivity analysis subordinated to a contrast. Commun Stat - Theory and Methods 45(15):4349–4364. https://doi.org/10.1080/03610926.2014.901369
Fuina JS, Pitangueira RLS, Penna SS (2011) A comparison of two microplane constitutive models for quasi-brittle materials. Appl Math Model 35(11):5326–5337. https://doi.org/10.1016/j.apm.2011.04.019
Gilquin L, Prieur C, Arnaud E, Monod H (2021) Iterative estimation of sobol’ indices based on replicated designs. Comput Appl Math 40:18. https://doi.org/10.1007/s40314-020-01402-5
Goda T (2021) A simple algorithm for global sensitivity analysis with shapley effects. Reliab Eng Syst Saf 213:107702. https://doi.org/10.1016/j.ress.2021.107702
Henkel R, Hoehndorf R, Kacprowski T, Knüpfer C, Liebermeister W, Waltemath D (2016) Notions of similarity for systems biology models. Briefings Bioinform 19(1):77–88. https://doi.org/10.1093/bib/bbw090
Heywood NI, Cheng DC-H (1984) Comparison of methods for predicting head loss in turbulent pipe flow of non-newtonian fluids. Trans Inst Measure Control 6(1):33–45. https://doi.org/10.1177/014233128400600105
Hübler C (2020) Global sensitivity analysis for medium-dimensional structural engineering problems using stochastic collocation. Reliab Eng Syst Saf 195:106749. https://doi.org/10.1016/j.ress.2019.106749
Jakeman JD, Eldred MS, Geraci G, Gorodetsky A (2020) Adaptive multi-index collocation for uncertainty quantification and sensitivity analysis. Int J Numer Methods Eng 121(6):1314–1343. https://doi.org/10.1002/nme.6268
Jamil R, Mujeebu MA (2019) Empirical relation between hazen-williams and darcy-weisbach equations for cold and hot water flow in plastic pipes. WATER 10, 104–114. https://doi.org/10.14294/WATER.2019.1
Kala Z (2019) Global sensitivity analysis of reliability of structural bridge system. Eng Struct 194:36–45. https://doi.org/10.1016/j.engstruct.2019.05.045
Larock BE, Jeppson RW, Watters GZ (2000) Hydraulics of Pipeline Systems. CRC Press, Boca Raton
Malena M, Portioli F, Gagliardo R, Tomaselli G, Cascini L, de Felice G (2019) Collapse mechanism analysis of historic masonry structures subjected to lateral loads: a comparison between continuous and discrete models. Comput Struct 220:14–31. https://doi.org/10.1016/j.compstruc.2019.04.005
McGuire W, Gallagher RH, Ziemian RD (2000) Matrix Structural Analysis, 2nd, edn. Wiley, New York
Mwasunda JA, Chacha CS, Stephano MA, Irunde JI (2022) Modelling cystic echinococcosis and bovine cysticercosis co-infections with optimal control. Computat Appl Math 41:342. https://doi.org/10.1007/s40314-022-02034-7
Ökten G, Liu Y (2021) Randomized quasi-monte carlo methods in global sensitivity analysis. Reliab Engi Syst Saf 210:107520. https://doi.org/10.1016/j.ress.2021.107520
Ooi EH, Ooi ET (2017) Mass transport in biological tissues: comparisons between single- and dual-porosity models in the context of saline-infused radiofrequency ablation. Appl Math Model 41:271–284. https://doi.org/10.1016/j.apm.2016.08.029
Papaioannou I, Straub D (2021) Variance-based reliability sensitivity analysis and the form \(\alpha \)-factors. Reliab Eng Syst Saf 210:107496. https://doi.org/10.1016/j.ress.2021.107496
Saltelli A (2002) Making best use of model evaluations to compute sensitivity indices. Comput Phys Commun 145(2):280–297. https://doi.org/10.1016/S0010-4655(02)00280-1
Saltelli A, Ratto M, Andres T, Campolongo F, Cariboni J, Gatelli D, Saisana M, Tarantola S (2007) Global Sensitivity Analysis. The Primer. John Wiley & Sons, Ltd, ???. https://doi.org/10.1002/9780470725184.ch4
Sobol’ IM (2003) Theorems and examples on high dimensional model representation. Reliability Engineering & System Safety 79(2), 187–193. https://doi.org/10.1016/S0951-8320(02)00229-6. SAMO 2001: Methodological advances and innovative applications of sensitivity analysis
Sobol’, I.M.: Global sensitivity indices for nonlinear mathematical models and their monte carlo estimates. Mathematics and Computers in Simulation 55(1), 271–280 (2001). https://doi.org/10.1016/S0378-4754(00)00270-6. The Second IMACS Seminar on Monte Carlo Methods
Sobol’ IM (1990) On sensitivity estimation for nonlinear mathematical models. Matematicheskoe Modelirovanie 2(1):112–118 ((in Russian))
Sobol’ IM, Tarantola S, Gatelli D, Kucherenko SS, Mauntz W (2007) Estimating the approximation error when fixing unessential factors in global sensitivity analysis. Reliab Eng Syst Saf 92(7):957–960. https://doi.org/10.1016/j.ress.2006.07.001
Tang K, Congedo PM, Abgrall R (2015) Sensitivity analysis using anchored anova expansion and high-order moments computation. Int J Numer Methods Eng 102(9):1554–1584. https://doi.org/10.1002/nme.4856
Vorel J, Marcon M, Cusatis G, Caner F, Di Luzio G, Wan-Wendner R (2021) A comparison of the state of the art models for constitutive modelling of concrete. Comput Struct 244:106426. https://doi.org/10.1016/j.compstruc.2020.106426
Yun W, Lu Z, Jiang X, Liu S (2016) An efficient method for estimating global sensitivity indices. Int J Numer Methods Eng 108(11):1275–1289. https://doi.org/10.1002/nme.5249
Zaicenco AG (2017) Sparse collocation method for global sensitivity analysis and calculation of statistics of solutions in spdes. Int J Numer Methods Eng 110(13):1247–1271. https://doi.org/10.1002/nme.5454
Zhang K, Lu Z, Cheng K, Wang L, Guo Y (2020) Global sensitivity analysis for multivariate output model and dynamic models. Reliab Eng Syst Saf 204:107195. https://doi.org/10.1016/j.ress.2020.107195
Zhou D, Han H, Ji T, Xu X (2016) Comparison of two models for human-structure interaction. Appl Math Model 40(5):3738–3748. https://doi.org/10.1016/j.apm.2015.10.049
Zhu X, Sudret B (2021) Global sensitivity analysis for stochastic simulators based on generalized lambda surrogate models. Reliab Eng Syst Saf 214:107815. https://doi.org/10.1016/j.ress.2021.107815
Acknowledgements
This research was partly supported by CNPq (Brazilian Research Council). These financial support are gratefully acknowledged.
Funding
Funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico (309846/2022-6).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
the authors declare that this work has no conflict of interest/competing interests.
Additional information
Communicated by Dan Goreac.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A relation to mean square error
Appendix A relation to mean square error
The total index with respect to model choice can be written as
Let us define the auxiliary random variable
If only two mathematical models \(f_1\), \(f_2\) are considered, we then have
Besides
Thus
and
Substitution of Eq. (A6) into Eq. (A3) gives
Consequently
The total index with respect to model choice can then be written as
However, we observe that the Root Mean Square (RMS) Error is defined as
For this reason we can write
or, alternatively,
This demonstrates Eq. (8).
Also note that
and
Thus
It is then possible to write
where
is the mean model, i.e. the point-wise mean between models \(f_1\) and \(f_2\). From these results we can also write
that demonstrates Eq. (9).
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Torii, A.J., Begnini, R., Kroetz, H.M. et al. Global sensitivity analysis for mathematical models comparison. Comp. Appl. Math. 42, 345 (2023). https://doi.org/10.1007/s40314-023-02484-7
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40314-023-02484-7