User Manual

Abstract

This documentation presents the features of the Uranie platform (based on Uranie v4.11.0), that is developed at CEA/DES. This platform is designed for uncertainty propagation, sensitivity analysis and computational code qualification, in a single software environment. It is largely based on the ROOT software (http://root.cern.ch/), an oriented-object framework that is designed and maintained by CERN, primarily used by the particle physicist to analyse the very large amount of data recorded at the LHC (Large Hadron Collider).

Commit:

61cebb3

Contents:

• 1. Overview: Uranie in a nutshell
  • 1.1. Introducing Uranie
    • 1.1.1. Uranie modules organisation
    • 1.1.2. External dependencies
  • 1.2. ROOT Environment
    • 1.2.1. Environment variables
    • 1.2.2. ROOT interpreter and runtime compiler
    • 1.2.3. Standard compilation
    • 1.2.4. Uranie namespace
    • 1.2.5. Important modifications going from ROOT v5 to ROOT v6
    • 1.2.6. References
  • 1.3. The Python Interface
    • 1.3.1. Python version: greater than 3.8
    • 1.3.2. Environment variables
    • 1.3.3. Using PyROOT
    • 1.3.4. The PyURANIE interface
    • 1.3.5. References
• 2. The DataServer module
  • 2.1. Introduction
  • 2.2. The TAttribute class
    • 2.2.1. Nature of the attribute
    • 2.2.2. List of variable information
    • 2.2.3. Examples of use of the class TAttribute
    • 2.2.4. Adding TAttribute when data are already available
    • 2.2.5. Introducing the TStochasticAttribute classes
  • 2.3. Data handling
    • 2.3.1. Main format of input/output
    • 2.3.2. Import data from an ASCII file
    • 2.3.3. Import data from a TNtuple/TDSNtuple/TTree
    • 2.3.4. Adding attributes to a TDataServer
    • 2.3.5. Merging two DataServer
    • 2.3.6. Pattern selection
    • 2.3.7. Export to an ASCII file
  • 2.4. Statistical treatments and operations
    • 2.4.1. Normalising the variable
    • 2.4.2. Computing the ranking
    • 2.4.3. Computing the elementary statistic
    • 2.4.4. The quantile computation
    • 2.4.5. Correlation matrix
  • 2.5. Visualisation dedicated to uncertainties
    • 2.5.1. Histogram
    • 2.5.2. Box-and-whisker(“boxplot”)
    • 2.5.3. CDF, CCDF curves
    • 2.5.4. Graph 2D with contour levels
    • 2.5.5. Graph 2D “profile”
    • 2.5.6. Graph 2D “Tufte”
    • 2.5.7. Graph 2D “pairs”
    • 2.5.8. Graph “CobWeb”
    • 2.5.9. QQ plot
    • 2.5.10. PP plot
  • 2.6. Combining these aspects: performing PCA
    • 2.6.1. PCA usage within Uranie
• 3. The Sampler module
  • 3.1. Introduction
  • 3.2. The Stochastic methods
    • 3.2.1. Introduction
    • 3.2.2. The main sampler classes
    • 3.2.3. Simple example
    • 3.2.4. TConstrLHS example
  • 3.3. Description of a correlation
    • 3.3.1. Imposing the correlation coefficients
    • 3.3.2. The copula classes
  • 3.4. QMC method
  • 3.5. The random fields
  • 3.6. OAT Design
    • 3.6.1. Introduction
    • 3.6.2. OAT design in Uranie
    • 3.6.3. TOATDesign
  • 3.7. The Vectorial Quantification method
• 4. The Launcher module
  • 4.1. Introduction.md
    • 4.1.1. Overview of a simple case
  • 4.2. External Code
    • 4.2.1. Code input and output files
  • 4.3. Distribution
    • 4.3.1. Multi-step remote launching to clusters
• 5. The Modeler Module
  • 5.1. Introduction
  • 5.2. The TLinearRegression Class
  • 5.3. Chaos polynomial expansion
    • 5.3.1. Nisp in a nutshell
  • 5.4. The kriging method
    • 5.4.1. Construction of a kriging model
    • 5.4.2. Usage of a Kriging model
• 6. The Sensitivity module
  • 6.1. The Sobol method
    • 6.1.1. Computation of Sobol’s sensitivity indices
• 7. The Optimizer module
  • 7.1. Function optimisation
    • 7.1.1. Rosenbrock function
• 8. The Relauncher module
  • 8.1. Introduction
  • 8.2. Relauncher abstraction levels
  • 8.3. TEval
    • 8.3.1. TCIntEval and TCJitEval
    • 8.3.2. TPythonEval
    • 8.3.3. TCodeEval
    • 8.3.4. Evaluation functions composition
  • 8.4. TRun
    • 8.4.1. TSequentialRun
    • 8.4.2. TThreadedRun
    • 8.4.3. TMpiRun
  • 8.5. TMaster
    • 8.5.1. Dealing with attributes
    • 8.5.2. TLauncher2
• 9. The Reoptimizer module
  • 9.1. Introduction
    • 9.1.1. local optimizer
    • 9.1.2. global optimizer
    • 9.1.3. Number of objectives
  • 9.2. Problem definition
    • 9.2.1. Objectives and Constraints
    • 9.2.2. Sizing of a hollow bar example problem
  • 9.3. Local solver
    • 9.3.1. TNlopt
    • 9.3.2. Solvers
  • 9.4. Global solver
    • 9.4.1. A step-by-step description of Vizir
    • 9.4.2. TVizir2 and TVizirIsland
    • 9.4.3. Solvers
• 10. The Metamodel Optimization module
  • 10.1. Introduction
  • 10.2. Efficient Global Optimization
    • 10.2.1. Introduction
    • 10.2.2. Problem definition
• 11. The Calibration module
  • 11.1. Introduction
    • 11.1.1. Distances and likelihoods used to compare observations and model predictions
  • 11.2. Calibration classes, distance and likelihood functions, observations and model
    • 11.2.1. General introduction to data and model definition
    • 11.2.2. Defining data, distance and likelihood functions
    • 11.2.3. Common methods of the calibration classes
    • 11.2.4. Use-case for this chapter
  • 11.3. Minimisation techniques
    • 11.3.1. Constructing the TMinimisation object
    • 11.3.2. Defining the TMinimisation properties
    • 11.3.3. Looking at the results
  • 11.4. Analytical linear Bayesian estimation
    • 11.4.1. Constructing the TLinearBayesian object
    • 11.4.2. Defining the TLinearBayesian properties
    • 11.4.3. Looking at the results
    • 11.4.4. Prediction of the variance
  • 11.5. Approximate Bayesian Computation techniques (ABC)
    • 11.5.1. Constructing the TRejectionABC object
    • 11.5.2. Defining the TRejectionABC properties
    • 11.5.3. Looking at the results
  • 11.6. Markov chain Monte Carlo approach
    • 11.6.1. Constructing the TMCMC object
    • 11.6.2. Defining the TMCMC properties
    • 11.6.3. Running the estimate, exporting and loading chains, and continuing the calculation
    • 11.6.4. Investigating the quality of the samples through diagnostics and plots
    • 11.6.5. Looking at the results
  • 11.7. CIRCE method
    • 11.7.1. Constructing the TCirce object
    • 11.7.2. Defining the TCirce properties
    • 11.7.3. Running the estimate
    • 11.7.4. Looking at the results
• 12. The Uncertainty modeler module
  • 12.1. Tests based on the Empirical Distribution Function (“EDF tests”)
• 13. Use-cases in C++
  • 13.1. Introduction
  • 13.2. Macros DataServer
    • 13.2.1. Macro “dataserverAttributes.C”
    • 13.2.2. Macro “dataserverMerge.C”
    • 13.2.3. Macro “dataserverLoadASCIIFilePasture.C”
    • 13.2.4. Macro “dataserverLoadASCIIFile.C”
    • 13.2.5. Macro “dataserverLoadASCIIFileYoungsModulus.C”
    • 13.2.6. Macro “dataserverLoadASCIIFileIonosphere.C”
    • 13.2.7. Macro “dataserverLoadASCIIFileCornell.C”
    • 13.2.8. Macro “dataserverComputeQuantile.C”
    • 13.2.9. Macro “dataserverGeyserStat.C”
    • 13.2.10. Macro “dataserverGeyserRank.C”
    • 13.2.11. Macro “dataserverNormaliseVector.C”
    • 13.2.12. Macro “dataserverComputeStatVector.C”
    • 13.2.13. Macro “dataserverComputeCorrelationMatrixVector.C”
    • 13.2.14. Macro “dataserverComputeQuantileVec.C”
    • 13.2.15. Macro “dataserverDrawQQPlot.C”
    • 13.2.16. Macro “dataserverDrawPPPlot.C”
    • 13.2.17. Macro “dataserverPCAExample.C”
  • 13.3. Macros Sampler
    • 13.3.1. Macro “samplingFlowrate.C”
    • 13.3.2. Macro “samplingLHS.C”
    • 13.3.3. Macro “samplingLHSCorrelation.C”
    • 13.3.4. Macro “samplingQMC.C”
    • 13.3.5. Macro “samplingBasicSampling.C”
    • 13.3.6. Macro “samplingOATRegular.C”
    • 13.3.7. Macro “samplingOATRandom.C”
    • 13.3.8. Macro “samplingOATMulti.C”
    • 13.3.9. Macro “samplingOATRange.C”
    • 13.3.10. Macro “samplingSpaceFilling.C”
    • 13.3.11. Macro “samplingMaxiMinLHSFromLHSGrid.C”
    • 13.3.12. Macro “samplingConstrLHSLinear.C”
    • 13.3.13. Macro “samplingConstrLHSEllipses.C”
    • 13.3.14. Macro “samplingSingularCorrelationCase.C”
  • 13.4. Macros Launcher
    • 13.4.1. Input/Output with vector and string: introduction to macros with multitype
    • 13.4.2. Macro “launchCodeReadMultiTypeRow.C”
  • 13.5. Macros Sensitivity
    • 13.5.1. Macro “sensitivityBrutForceMethodFlowrate.C”
    • 13.5.2. Macro “sensitivityFiniteDifferencesFunctionFlowrate.C”
    • 13.5.3. Macro “sensitivityDataBaseFlowrate.C”
    • 13.5.4. Macro “sensitivityFASTFunctionFlowrate.C”
    • 13.5.5. Macro “sensitivityRBDFunctionFlowrate.C”
    • 13.5.6. Macro “sensitivityMorrisFunctionFlowrate.C”
    • 13.5.7. Macro “sensitivityMorrisFunctionFlowrateRunner.C”
    • 13.5.8. Macro “sensitivityRegressionFunctionFlowrate.C”
    • 13.5.9. Macro “sensitivitySobolFunctionFlowrate.C”
    • 13.5.10. Macro “sensitivitySobolFunctionFlowrateRunner.C”
    • 13.5.11. Macro “sensitivityRegressionLeveLE.C”
    • 13.5.12. Macro “sensitivitySobolLeveLE.C”
    • 13.5.13. Macro “sensitivitySobolRe-estimation.C”
    • 13.5.14. Macro “sensitivitySobolWithData.C”
    • 13.5.15. Macro “sensitivitySobolLoadFile.C”
    • 13.5.16. Macro “sensitivityJohnsonRWFunctionFlowrate.C”
    • 13.5.17. Macro “sensitivityJohnsonRWCorrelatedFunctionFlowrate.C”
    • 13.5.18. Macro “sensitivityJohnsonRWJustCorrelationFakeFlowrate.C”
    • 13.5.19. Macro “sensitivityHSICFunctionFlowrate.C”
    • 13.5.20. Macro “sensitivitySobolRankFunctionFlowrate.C”
  • 13.6. Macros Modeler
    • 13.6.1. Macro “modelerCornellLinearRegression.C”
    • 13.6.2. Macro “modelerFlowrateLinearRegression.C”
    • 13.6.3. Macro “modelerFlowrateMultiLinearRegression.C”
    • 13.6.4. Macro “modelerFlowrateNeuralNetworks.C”
    • 13.6.5. Macro “modelerFlowrateNeuralNetworksLoadingPMML.C”
    • 13.6.6. Macro “modelerClassificationNeuralNetworks.C”
    • 13.6.7. Macro “modelerFlowratePolynChaosRegression.C”
    • 13.6.8. Macro “modelerFlowratePolynChaosIntegration.C”
    • 13.6.9. Macro “modelerbuildSimpleGP.C”
    • 13.6.10. Macro “modelerbuildGPInitPoint.C”
    • 13.6.11. Macro “modelerbuildGPWithAPriori.C”
    • 13.6.12. Macro “modelerbuildSimpleGPEstim.C”
    • 13.6.13. Macro “modelerbuildSimpleGPEstimWithCov.C”
    • 13.6.14. Macro “modelerTestKriging.C”
  • 13.7. Macros Relauncher
    • 13.7.1. Macro “relauncherFunctionFlowrateCInt.C”
    • 13.7.2. Macro “relauncherFunctionFlowrateCJit.C”
    • 13.7.3. Macro “relauncherCJitFunctionThreadTest.C”
    • 13.7.4. Macro “relauncherCodeFlowrateSequential.C”
    • 13.7.5. Macro “relauncherCodeFlowrateSequential_ConstantVar.C”
    • 13.7.6. Macro “relauncherCodeFlowrateThreaded.C”
    • 13.7.7. Macro “relauncherCodeFlowrateMPI.C”
    • 13.7.8. Macro “relauncherCodeFlowrateMpiStandalone.C”
    • 13.7.9. Macro “relauncherCodeFlowrateSequentialFailure.C”
    • 13.7.10. Macro “relauncherCodeMultiTypeKey.C”
    • 13.7.11. Macro “relauncherCodeMultiTypeKeyEmptyVectors.C”
    • 13.7.12. Macro “relauncherCodeMultiTypeKeyEmptyVectorsAsFailure.C”
    • 13.7.13. Macro “relauncherCodeReadMultiType.C”
    • 13.7.14. Macro “relauncherComposeMultitypeAndReadMultiType.C”
    • 13.7.15. Macro “relauncherCodeFlowrateSequential_TemporaryVar.C”
  • 13.8. Macros Reoptimizer
    • 13.8.1. Macro “reoptimizeHollowBarCode.C”
    • 13.8.2. Macro “reoptimizeHollowBarCodeMultiStart.C”
    • 13.8.3. Macro “reoptimizeHollowBarCodevizir.C”
    • 13.8.4. Macro “reoptimizeHollowBarVizirMoead.C”
    • 13.8.5. Macro “reoptimizeHollowBarVizirSplitRuns.C”
    • 13.8.6. Macro “reoptimizeZoneBiSubMpi.C”
    • 13.8.7. Macro “reoptimizeZoneBiFunMpi.C”
  • 13.9. Macros MetaModelOptim
    • 13.9.1. Macro “metamodoptEgoHimmel.C”
  • 13.10. Macros Calibration
    • 13.10.1. Macro “calibrationMinimisationFlowrate1D.C”
    • 13.10.2. Macro “calibrationMinimisationFlowrate2DVizir.C”
    • 13.10.3. Macro “calibrationLinBayesFlowrate1D.C”
    • 13.10.4. Macro “calibrationRejectionABCFlowrate1D.C”
    • 13.10.5. Macro “calibrationMCMCFlowrate1D.C”
    • 13.10.6. Macro “calibrationMCMCLinReg.C”
    • 13.10.7. Macro “calibrationCirce.C”
  • 13.11. Macros UncertModeler
    • 13.11.1. Macro “uncertModelerTestsYoungsModulus.C”

[App13]

W. Appel. Probabilité pour les non probabilistes. H & K, Paris, 2013.

[Arn10]

G. Arnaud. Manuel d'utilisation de Vizir distribué v2.0. Technical Report, CEA, SFME/LGLS/RT/10-001/A, 2010.

[Bla17]

J-B. Blanchard. Methodological reference guide for uranie v3.11.0. Technical Report, CEA, DEN/DANS/DM2S/STMF/LGLS/RT/17-006/A, 2017. Updated with every new release.

[BR97]

R. Brun and F. Rademakers. ROOT: An object oriented data analysis framework. Nucl. Instrum. Meth., A389:81–86, 1997. doi:10.1016/S0168-9002(97)00048-X.

[FL02]

Michael Feathers and B Lepilleur. Cppunit cookbook. 2002.

[FJ05]

Matteo Frigo and Steven G. Johnson. The design and implementation of FFTW3. Proceedings of the IEEE, 93(2):216–231, 2005. Special issue on “Program Generation, Optimization, and Platform Adaptation”.

[GFB+04]

Edgar Gabriel, Graham E. Fagg, George Bosilca, Thara Angskun, Jack J. Dongarra, Jeffrey M. Squyres, Vishal Sahay, Prabhanjan Kambadur, Brian Barrett, Andrew Lumsdaine, Ralph H. Castain, David J. Daniel, Richard L. Graham, and Timothy S. Woodall. Open MPI: goals, concept, and design of a next generation MPI implementation. In Proceedings, 11th European PVM/MPI Users' Group Meeting, 97–104. Budapest, Hungary, September 2004.

[GRG+96]

Andrew Gelman, Gareth O Roberts, Walter R Gilks, and others. Efficient metropolis jumping rules. Bayesian statistics, 5(599-608):42, 1996.

[Gil09]

N. Gilardi. Interface python pour la plate-forme uranie. Technical Report, CEA, SFME/LGLS/RT/09-015/A, 2009.

[Hal64]

J. H. Halton. Algorithm 247: radical-inverse quasi-random point sequence. Commun. ACM, 7(12):701–702, December 1964. URL: http://doi.acm.org/10.1145/355588.365104, doi:10.1145/355588.365104.

[HD02]

J. C. Helton and F. J. Davis. Illustration of sampling-based methods for uncertainty and sensitivity analysis. Risk Analysis, 22(3):591–622, 2002. URL: http://dx.doi.org/10.1111/0272-4332.00041, doi:10.1111/0272-4332.00041.

[IC82]

R. L. Iman and W. J. Conover. A distribution-free approach to inducing rank correlation among input variables. Communications in Statistics - Simulation and Computation, 11(3):311–334, 1982. doi:10.1080/03610918208812265.

[Joh]

Steven G. Johnson. The nlopt nonlinear-optimization package. http://ab-initio.mit.edu/nlopt.

[Jol11]

Ian Jolliffe. Principal component analysis. Springer, 2011.

[JSW98]

Donald R Jones, Matthias Schonlau, and William J Welch. Efficient global optimization of expensive black-box functions. Journal of Global optimization, 13(4):455–492, 1998.

[MH07]

Ken Martin and Bill Hoffman. An open source approach to developing software in a small organization. Ieee Software, 2007.

[MBC00]

M. D. McKay, R. J. Beckman, and W. J. Conover. A comparison of three methods for selecting values of input variables in the analysis of output from a computer code. Technometrics, 42(1):55–61, February 2000. URL: http://dx.doi.org/10.2307/1271432, doi:10.2307/1271432.

[MOHW07]

J. C. Meza, R. A. Oliva, P. D. Hough, and P. J. Williams. Opt++: an object-oriented toolkit for nonlinear optimization. ACM Transactions on Mathematical Software, June 2007. URL: http://doi.acm.org/10.1145/1236463.1236467, doi:10.1145/1236463.1236467.

[Nvi11]

CUDA Nvidia. Nvidia cuda c programming guide. Nvidia Corporation, 120(18):8, 2011.

[Pet01]

K. Petras. Fast calculation of coefficients in the smolyak algorithm. Numerical Algorithms, 26(2):93–109, 2001. URL: http://dx.doi.org/10.1023/A:1016676624575, doi:10.1023/A:1016676624575.

[RGG+97]

Gareth O Roberts, Andrew Gelman, Walter R Gilks, and others. Weak convergence and optimal scaling of random walk metropolis algorithms. The annals of applied probability, 7(1):110–120, 1997.

[Sobol67]

I.M Sobol'. On the distribution of points in a cube and the approximate evaluation of integrals. USSR Computational Mathematics and Mathematical Physics, 7(4):86 – 112, 1967. URL: http://www.sciencedirect.com/science/article/pii/0041555367901449, doi:http://dx.doi.org/10.1016/0041-5553(67)90144-9.

[Wil13]

Richard David Wilkinson. Approximate bayesian computation (abc) gives exact results under the assumption of model error. Statistical applications in genetics and molecular biology, 12(2):129–141, 2013.

[Wor87]

Brian A Worley. Deterministic uncertainty analysis. Technical Report, Oak Ridge National Lab., TN (USA), 1987.