The SUMO toolbox  2017a
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Pages
Class List
Here are the classes, structs, unions and interfaces with brief descriptions:
[detail level 123456]
 Niminds
 CAggregateObjectiveAggregates several candidateRankers into one score (using a weighted average)
 CAICCalculates Akaike's information criteria (AIC)
 CANNFactoryThis class is responsible for generating ANN models based on the Matlab ANN toolbox
 CANNGenerationObservableThis observable monitors how often each learning rule is used in a population of ANN models
 CANNModelConstructs a new feedforward Artificial Neural Network (ANN)
 CAssignLabelsCan be used to define labels on a continuous interval
 CAudzeEglaisDistanceSelects the candidates with the highest AudzeEglais distance to existing points
 CBasicGaussianProcessA kriging surrogate model (also known as a Gaussian Process)
 CBatchObservableThis observable works on a set of models, instead of on a single model
 CBlindKrigingA blind kriging surrogate model
 CBoxBehnkenDesignChoose an initial sampleset according to a Box-Behnken design
 CCandidateGeneratorAn abstract class that provides an easy and convenient interface for generating candidates, either by subclassing from this one, or by just calling it directly, through a function
 CCandidateRankerAn interface that allows the object to score a set of candidates according to its own system
 CCategoriesObservableThis observable monitors how often each model type occurs in a population
 CCentralCompositeDesignChoose an initial sampleset according to a central composite design
 CCloneable
 CClosenessThresholdSelects a set of new samples from the candidates by selecting the best scoring n candidates, but avoiding samples that lie too close to each other
 CCMAESOptimizerWrapper around the CMA-ES optimization algorithm
 CCoKrigingA cokriging surrogate model
 CCombinedDesignWrap 2 different Initial Designs Together
 CCombinedModelBuilderRuns multiple modelbuilders in sequence
 CCombinedSequentialDesignJust a class to wrap together 2 different sample selectors
 CCombiOptimizerWrapper allowing usage of several optimizer together
 CComplexWrapperA utility class to wrap 2 models for a complex output
 CConstraintAbstract base class representing a constraint
 CConstraintManagerManages several constraint classes
 CCrossCornerDesignCross Corner design - just generate 2 points at [-1,...-1] and [1,...,1]
 CCrossValidationThis measure uses k-fold crossvalidation to gauge the accuracy of the model
 CcrowdednessCalculates the crowdedness at a given design x or in this case, for all designs in 'points')
 CDataModelWraps a dataset in a SUMO Model object
 CDataModifierA DataModifier modifies data coming from a simulator, for example introducing noise, or taking the log
 CDatasetDesignRead an initial design from a dataset file dataset can also be specified using the id from a simulator xml
 CDatasetDirectDataSourceA fast data source which reads data from files and doesn't use java code
 CDegreesConstruct a degree class
 CDelaunayCandidateGeneratorGenerates samples based on a Delaunay triangulation
 CDelaunayMergerDelaunaydMerger performs a merging of the candidates for each simplex (generated using a Delaunay triangulation)
 CdelaunayVolumeCompute the volume of the delaunay triangle containing each candidate
 CDIANNEFactoryA very simple class that generates ELMModel objects
 CDIANNEModelConstructs a new Extreme Learning Machine (ELM)
 CDifferentialEvolutionDifferential Evolution (DE) algorithm
 CdimensionDistanceCalculate the non-collapsing factor of the candidates
 CDirectOptimizerWrapper around the DIRECT optimization algorithm
 CDiscreteOptimizerDiscrete optimizer
 CDistanceTODO
 CEGOModelBuilderUses the Efficient Global Optimization (EGO) algorithm to optimize in hyperparameter space
 CELMFactoryA very simple class that generates ELMModel objects
 CELMModelConstructs a new Extreme Learning Machine (ELM)
 CEmptyDesignChoose no samples at all
 CEmptyModelEmpty zero-model
 CEmptyModelBuilderGenerates a zero model by default
 CEmptySequentialDesignThis sample selector always return an empty selection
 CEnsembleFactoryResponsible for generating weighted ensemble models
 CEnsembleModelConstructs a basic weighted ensemble object
 CEnsembleObservableThis observable tracks the composition of the best performing ensemble
 CEnsembleRankerCalculates the expected improvement statistical infill criterion
 CEuclideanDistanceTODO
 CEureqaFactoryModelFactory for the Eureqa Symbolic Regression Tool
 CEureqaModelWrapper for the Eureqa Symbolic Regression Tool
 CEvaluationTimeMeasureMeasures the evaluation time for a model
 CexpectedImprovementCalculates the expected improvement statistical infill criterion
 CexpectedImprovementEuclideanCalculates the expected improvement statistical infill criterion
 CexpectedImprovementHypervolumeCalculates the hypervolume-based expected improvement statistical infill criterion
 CExpressionModelWraps any matlab expression in a SUMO Model object
 CFactorialDesignFactorial design
 CFailureIntroduces failed simulations to the data (NaN's)
 CFANNFactoryThis class is responsible for generating neural network models as implemented in the Fast Artificial Neural Network Libary (FANN) http://leenissen.dk/fann/
 CFANNModelConstructs a new neural network backed by the FANN library
 CFLOLASampleRankerA class that rates samples according to the Fuzzy-LOLA non-linearity criterion
 CFractionalDistanceTODO
 CfuzzySpaceFillingSpace-filling score based on fuzzy logic
 CGaussianProcessFactoryThis class is responsible for generating Gaussian Process Models
 CGaussianProcessModelRepresents a Gaussian Process surrogate model
 CGeneticFactoryFactories that support the Genetic Model Builder must derive from this class
 CGeneticModelBuilderUses a Genetic Algorithm (GA) to select the best model parameters
 CgExpectedImprovementCalculates the Generalized Expected Improvement
 CgProbabilityOfFeasibilityCalculates the generalized probability of feasibility for a point
 Chandle
 CHeterogeneousFactoryThis is a meta-Factory that wraps other factories as part of the heterogeneous evolution for model type selection
 CHilbertCurveChoose points so that they form a Hilbert curve in 2 dimensions Ref: http://blogs.mathworks.com/steve/2012/01/25/generating-hilbert-curves/
 CHillClimberOptimizerSimple hill climbing algorithm, starting from a large initial population
 ChypervolumePoIQuantifies the PoI with the hypervolume contribution (for MOSBO problems)
 CInitialDesignBase class for all initial design generators
 CinputParser
 CinputParserSUMOSUMO version of an inputParser
 CInterpolationFactoryA very simple class that generates InterpolationModel objects
 CInterpolationModelConstructs an interpolation based on Matlabs griddata (to allow for scattered data)
 CknowledgeGradientCalculates the knowledge gradient statistical infill criterion
 CKrigingA kriging surrogate model
 CKrigingFactoryThis class is responsible for generating Kriging Models
 CkushnerCalculates the kushner criterion
 CLatinHypercubeDesignChoose an initial sampleset in such a way that they form a latin hypercube
 CLeaveOneOutThis measure performs Leave-One-Out crossvalidation to gauge the accuracy of the model
 CLevelPlotThe class holds the data necessary to generate LevelPlots
 CLHDOptimizerA quasi-LHD optimizer
 CLinearConstraintImplementation of a linear constraint
 CLocalPatternSearchOptimizer which generates quasi-latin hypercubes through genetic algorithm optimization
 CLogTransformTakes the logarithm of the data
 CLOLASampleRankerA class that rates samples according to the LOLA non-linearity criterion
 CLOLAVoronoiSequentialDesignUses the LOLASampleRanker and VoronoiSampleRanker to balance exploration (searching the input space) and exploitation (focussing on regions of non-linearity)
 ClowerConfidenceBoundCalculates the lower confidence bound
 ClrmThe LRM candidate ranker
 CLRMMeasureReturn a score based on how much a model approaches a linear fit (the more linear the lower the score) LRM: Linear Reference Model
 CLSSVMFactoryThe class serves as a kind of base class for LS-SVM based modelers, it takes care of parsing and holding the basic configuration options that any LSSVM modeler needs
 CLSSVMModelConstructs a new Least Squares Support Vector Machine (SVM) model using the given configuration
 CMatlabDirectDataSourceA fast data source for Matlab functions which doesn't use java code
 CMatlabGAWrapper around the matlab optimizers
 CMatlabOptimizerWrapper around the matlab optimizers
 CMatlabPatternSearchWrapper around the matlab optimizers
 CMatlabSimAnnealingWrapper around the matlab optimizers
 CmaximinDistanceSelects the candidates with the highest maximin distance to existing points
 CMaximinDistanceConstraintImplementation of a maximin distance constraint
 CmaximinManhattanDistanceSelects the candidates with the highest maximin distance to existing points
 CMeasureAbstract base class for a measure
 CMergeCriterionImplement this interface if you want to be able to select a set of samples that have to be evaluated from a set of candidates, based on one or more rankings provided by other objects
 CMinMaxThis measure enforces the minimum and maximum that is defined for the output
 CModelConstructs an abstract Model object
 CModelBuilderAdaptive model builder base class
 CModelContainerMerges one or more sumo Models together in one Model
 CModelDifferenceEstimate the accuracy of the model by comparing it with all the other models and assuming that, when models are similar, the algorithm is converging to a stable model, which should be the correct one
 CmodelDifferenceCalculates the difference between the last nrModels sumo models
 CmodelEvaluateSimple function that evaluates the model directly
 CModelFactoryBase class for all model factories
 CModelGridManagerCreate a new empty model grid manager which can be used to store & get the evaluated grid of a model
 CModelInterfaceThe model interface provides the set of abstract functions that must be supported by any model, be it real or wrapper
 CmodelVarianceCalculates the prediction variance
 CNLOPTOptimizerWrapper for the NLOPT optimization library
 CNoiseIntroduces noise to the data
 CNonlinearConstraintNonlinear constraint, accepts an arbitrary function handle that implements the constraint
 CObservableAn Observable is an object that is able to extract model parameter values from a model so they can be monitored (plotted) during the modeling process
 COptimizeCriterionThis sample selector selects one or more samples that optimizes a certain candidateRanker
 COptimizerAbstract base class for an optimizer
 COptimizerModelBuilderOptimizes the model parameters using one the Optimizers available in src/matalb/tools/Optimizers
 COutlierIntroduces outliers to the data
 COutputFilterWrapperThis class wraps another model, hiding one or more outputs
 CParetoFrontUpdates the cells (integral bounds) for the pareto front
 CParetoModelBuilderUses a Multiobjective GA to select the best model parameters
 CPCTOptimizerAn optimizer which handles multiple points in parallel and makes sure they are far away from each other (as much as possible) using ClosenessThreshold
 CphiDistanceSelects the candidates with the highest maximin distance to existing points
 CPipelineSequentialDesignThe PipeLineSequentialDesign generates a number of points using a CandidateGenerator, then evaluates all these points on one or more criteria (CandidateRanker), and then a MergeCriterion is used to merge these scores and select the samples from them
 CPolynomialFactoryThis class is responsible for generating Polynomial Models
 CPolynomialModelConstruct a `PolynomialModel' object
 CpredictiveEntropyCalculates the expected improvement statistical infill criterion
 CprobabilityOfFeasibilityCalculates the probability of feasibility for a point
 CprobabilityOfImprovementCalculates the probability of improvement for a point
 CProjectedDistanceGridCandidateGeneratorGenerates all the local optima for the projected distance criterion exactly, by using the inherent properties of the surface
 CProjectedDistanceGridOptimizerOptimizes the points generated by the ProjectedDistanceGridCandidateGenerator towards a particular criterion
 CProjectedDistanceThresholdA criterion that ranks all points according to their projected distance score
 CProjectedThresholdRandomCandidateGeneratorTODO
 CpsiDistanceSelects the candidates with the highest psi distance to existing points
 CPSOOptimizerWrapper around the Another PSO library (Particle Swarm Optimization)
 CPSOtOptimizerWrapper around the PSOt library (Particle Swarm Optimization)
 CQuasiRandomDesignGenerates a space-filling initial design by generating the first of a set of quasi-random numbers
 CRandomCandidateGeneratorGenerates candidates from the uniform distribution
 CRandomDesignChoose samples randomly
 CRandomModelBuilderGenerate random models, usefull as a baseline benchmark
 CRandomSequentialDesignChooses datapoints in random locations
 CRandomZoomCandidateGeneratorGenerates candidates from the uniform distribution near existing maximin samples
 CRationalFactoryThis class generates Rational models
 CRationalModelConstruct a `RationalModel' object
 CrationalPoleSupressionPromotes points that makes the denominator zero, thus finding poles
 CRBFFactoryThis class is responsible for generating RBF models
 CRBFModelConstructs a radial basis function model object
 CRBFNNFactoryThis class is responsible for generating Radial Basis Function Neural Networks as implemented in the Matlab NN toolbox
 CRBFNNModelConstructs a new Radial Basis Function Neural Network (RBFNN)
 CRunnable
 CSampleErrorThis measure simply compares the values of the model at the locations of all the samples used for constructing it with the actual values of these samples
 CSampleManagerThis class processes new samples and stores them, so that they can later be requested in both simulator and toolbox space
 CSampleRankerRanks existing samples
 CScaleByConstantScales the data by a constant
 CSDPDistanceTODO
 CSensitivityCrossValidationThis measure uses k-fold crossvalidation to gauge the accuracy of the model
 CSeparateCriteriaMerge criterion that select the top samples of each measure independently, in a cyclic fashion
 CSequentialDesignAn abstract base class for every SequentialDesign
 CSequentialDesignTypesEnumeration class to represent whether a sequential design is input, output or model based
 CSequentialInitialDesignUses a space filling sequential design to create an initial design
 CSequentialModelBuilderAdaptive model builder subclass that builds models sequentially
 CSimpleObservableThis is the simplest form of observable
 CSliceSampleFactoryBase class for all model factories
 CSobolHelper class to estimate sobol indices (any order)
 CSplineFactoryThis class is responsible for generating Smoothing Spline models (1D and 2D only), based on the Matlab splines toolbox
 CSplineModelConstructs a new Spline model based on the smoothing spline implementation from the Matlab Splines toolbox
 CSQPLabOptimizerWrapper around the SQPLab optimization package
 CSUMOThe main class of the SUMO Toolbox
 CSVMFactoryThe class serves as a kind of base class for SVM based modelers, it takes care of parsing and holding the basic configuration options that any SVM modeler needs
 CSVMModelConstructs a new Support Vector Machine (SVM) model using the given configuration
 CTestEngineRuns test cases (regression tests)
 CTestMinimumThis measure is used for validition purpose of the optimization framework The true minimum (function value!) is compared to the current minimum found
 CTimeMeasureMeasure based on the current time
 CTPLatinHypercubeDesignChoose an initial sampleset in such a way that they form a latin hypercube
 CTrainingTimeMeasureMeasures the training time of a model
 CTriangulationA handle class that holds and updates a triangulation
 CValidationSetThis measure uses a set of validation samples, not used for construction of the model, to estimate the accuracy of the model
 CVoronoiEdgeTraversalSequentialDesignSelects points along the Voronoi edges of the Tessellation of existing samples
 CVoronoiSampleRankerRanks samples based on a voronoi diagram
 Cwb1Calculates the threshold-bounded extreme
 Cwb2Locates the regional extreme
 CWeightedAverageWeightedAverage performs a weighted averaged merging of the different scores
 CWeightedAverageOnePerSimplexWeighted averaging of the candidateRankers
 CWeightedAverageSnapToBorderWeightedAverage performs a weighted averaged merging of the different scores
 CwExpectedImprovementWeighted expected improvement
 CWrappedModelThis model is a wrapper around another model