KerMor  0.9
Model order reduction for nonlinear dynamical systems and nonlinear approximation
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Pages
Properties with level "Critical"

Property level critical

Member approx.algorithms.AAdaptiveBase.MaxRelErr
Member approx.algorithms.ABase.ExpConfig
Without this setting this algorithm makes little sense.
Member approx.algorithms.Componentwise.CoeffConfig
Without this setting this algorithm makes little sense.
Member approx.KernelApprox.Algorithm
The approximation strategy for kernel expansions is essential.
Member approx.TPWLApprox.Beta
Member data.selection.EpsSelector.EpsRad
Determines how many training samples are taken. Value MUST be set taking into account the full system's dimension or at least the bounding box of the samples!
Member data.selection.LinspaceSelector.Size
The amount of approximation training data to take.
Member data.selection.TimeSelector.Size
Determines the (maximum) size of the approximation training data samples to take.
Member DPCMDemoClass.Prop1

This is a very critical property!

This is a very critical property!

Member dscomponents.ACoreFun.fDim
Not setting this value in inheriting classes may cause errors in some KerMor algorithms.
Member dscomponents.ACoreFun.TimeDependent
Not setting this value in implementing subclasses causes KerMor's ODE solvers to (possibly) produce wrong results due to wrong assumptions on the time dependence of the core function.
Member dscomponents.ACoreFun.xDim
Not setting this value in inheriting classes may cause errors in some KerMor algorithms.
Member dscomponents.AInputConv.TimeDependent
Not setting this value in implementing subclasses causes KerMor's ODE solvers to (possibly) produce wrong results due to wrong assumptions on the time dependence of the core function.
Member dscomponents.AOutputConv.TimeDependent
Some output conversion matrices are time dependent. This property must be set to the correct value in order for the output conversion to work correctly.
Member general.AffParamMatrix.Matrices
Member general.POD.UseSVDS
Allows for dramatic reduction of computational costs for the price of less accuracy. See the matlab docs for svds and svd.
Member general.regression.BaseScalarSVR.Lambda
Overly regularized functions may not approximate the data correctly, while small \(\lambda\) lead to high coefficient values.
Member general.regression.KernelLS.CGTol
The error tolerance for the pcg method.
Member general.regression.ScalarEpsSVR.Eps
Determines the precision of the approximation.
Member general.regression.ScalarEpsSVR_SMO.Eps
Influences the precision of the result.
Member general.regression.ScalarEpsSVR_SMO.StopEps
Determines the precision of the solution. This eps is used within the stopping criterion introduced in SHS11.
Member general.regression.ScalarNuSVR.nu
Too low \(\nu\) parameters might result in too few support vectors and thus large \(epsilon\). Is closely connected to the C property.
Member kernels.ARBFKernel.Gamma
Greatly influences the kernels behaviour.
Member kernels.BaseKernel.G
If a custom norm is used (i.e. after subspace projection) tihs must be set in order to obtain correct evaluations.
Member kernels.BellFunction.r0
This value is essential for any bell function.
Member kernels.InvMultiquadrics.beta
Greatly influences the kernels behaviour.
Member kernels.InvMultiquadrics.c
Greatly influences the kernels behaviour.
Member kernels.KernelExpansion.Kernel
Correct choice of the system kernel greatly influences the function behaviour.
Member kernels.ParamTimeKernelExpansion.ParamKernel
Correct choice of the system kernel greatly influences the function behaviour.
Member kernels.ParamTimeKernelExpansion.TimeKernel
Correct choice of the time kernel greatly influences the function behaviour.
Member kernels.PolyKernel.Degree
Greatly influences the kernels behaviour.
Member kernels.SigmoidKernel.kappa
Greatly influences the kernels behaviour.
Member kernels.SigmoidKernel.nu
Greatly influences the kernels behaviour.
Member models.BaseFirstOrderSystem.DependentParamIndices
Using more \(\mu\) parameter vector elements in the kernel approximations than actually required in the core function \(f\) introduces an additional dependency of the nonlinearity on extra parameters which is not given in the full model's core function.
Member models.BaseFirstOrderSystem.MaxTimestep
Too large time-steps might violate e.g. CFL conditions and render the solution false and useless.
Member models.BaseFirstOrderSystem.x0
The initial value greatly influences the simulation results.
Member models.BaseFullModel.Approx
The approximation technique is critical for the quality of the reduced model.
Member models.BaseFullModel.SpaceReducer
The subspace computation strategy is critical for the quality of the projection subspace
Member models.BaseModel.dt
Member models.BaseModel.System
No simulations without dynamical system.
Member models.pcd.BasePCDSystem.h
Determines the spatial resolution of the model.
Member models.pcdi.BasePCDISystem.h
Determines the spatial resolution of the model.
Member sampling.ManualSampler.Samples
If manual sampling is used, this property builds the basic set of parameter samples to be used for offline computations and hence must be chosen with maximum care.
Member sampling.RandomSampler.Samples
Determines how many parameter samples are taken and thus directly the offline computation time and model approximation quality.
Member solvers.BaseSolver.MaxStep
Too large steps due to high time-step distances in the passed times vector \(t\) may lead to errorneous results. This property limits the maximum time-step size used in the implementations. Set to [] in order to rely on the times \(t\).
Member solvers.MLode15i.AbsTol
Absolute error tolerances require adoption to the current situation and scale of computations.
Member solvers.MLode15i.RelTol
The correct relative tolerance guarantees the right number of significant figures.
Member solvers.MLWrapper.MLSolver
The correct underlying MatLab builtin solver can make the difference.
Member spacereduction.PODGreedy.Eps
Ultimately determines the projection error on all training trajectories.