The SUMO toolbox  2018a
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Pages
Public Member Functions | Protected Member Functions | List of all members
GaussianProcessFactory Class Reference

This class is responsible for generating Gaussian Process Models. More...

Inheritance diagram for GaussianProcessFactory:
Inheritance graph
[legend]

Public Member Functions

function GaussianProcessFactory (var varargin)
 Class constructor. More...
 
function getMeanFunction ()
 Returns the set of mean functions. More...
 
function getCovarianceFunction ()
 Returns the set of covariance functions. More...
 
function getLikelihoodFunction ()
 Returns the set of likelihood functions. More...
 
function getBackend ()
 Returns a user-friendly description of this factory. More...
 
function supportsComplexData ()
 Defines whether this factory supports complex data. More...
 
function supportsMultipleOutputs ()
 Defines whether this factory supports multiple outputs. More...
 
function getBounds ()
 Returns bounds on the hyperparameters. More...
 
function createRandomModel ()
 creates a random model More...
 
function createInitialModels (var number, var wantModels)
 creates a set of initial models More...
 
function createModel (var varargin)
 creates one model More...
 
function getObservables ()
 Defines the supported observables of this surrogate model. More...
 
function getModelType ()
 Returns the supported surrogate model type. More...
 
function mutation (var parents, var options, var nvars, var FitnessFcn, var state, var thisScore, var thisPopulation)
 Mutation operator for a Genetic Algoritm. More...
 
function crossover (var parents, var options, var nvars, var FitnessFcn, var unused, var thisPopulation)
 Crossover operator for a Genetic Algoritm. More...
 
function isSliceSamplingEnabled ()
 Performs elliptical slice sampling. More...
 
function getConstraintFcn ()
 TODO. More...
 
function getCreationFcn ()
 TODO. More...
 
function getCrossoverFcn ()
 TODO. More...
 
function getMutationFcn ()
 TODO. More...
 
function isCustomMode ()
 TODO. More...
 
function getBatchObservables ()
 TODO. More...
 
function getIndividualSize ()
 TODO. More...
 
function setSamples (var samples, var values)
 TODO. More...
 
function getIntegerParameters ()
 MATLAB GA supports handling parameters as integers, override this function in derived factories when parameters must be handled as integers. More...
 
function getBasicBatchObservables ()
 TODO. More...
 
function createInitialPopulation (var GenomeLength, var FitnessFcn, var options)
 TODO. More...
 
function wrapFunctions ()
 TODO. More...
 
function getSamples ()
 TODO. More...
 
function getDimensions ()
 TODO. More...
 
function getParallelMode ()
 TODO. More...
 
function getMode ()
 TODO. More...
 

Protected Member Functions

function sliceSample (var model)
 Given an individual kriging model, sample the likelihood using slice sampling. More...
 
function performSliceSampling (var hp0, var likelihood)
 Performs standard slice sampling. More...
 

Detailed Description

This class is responsible for generating Gaussian Process Models.

Uses the Gaussian Processes for Machine Learning (GPML) toolbox

Constructor & Destructor Documentation

function GaussianProcessFactory ( var  varargin)
inline

Class constructor.

Initializes the factory

Returns
instance of the factory

Member Function Documentation

function createInitialModels ( var  number,
var  wantModels 
)

creates a set of initial models

Return 'number' individuals.

Return values
modelsset of models

If wantModels is false only return a parameter matrix where each row uniquely represents one model. If wantModels is true an array of model objects is returned.

function createInitialPopulation ( var  GenomeLength,
var  FitnessFcn,
var  options 
)
inherited

TODO.

A function that creates an initial population.

The input arguments to the function are Genomelength : Number of independent variables for the fitness function FitnessFcn : Fitness function options : Options structure The function returns Population, the initial population for the genetic algorithm.

function createModel ( var  varargin)

creates one model

Given an individual representing a model, return a real model.

Parameters
vararginparameters of model
Return values
modelone model
function createRandomModel ( )
inline

creates a random model

Return values
modela random model
function crossover ( var  parents,
var  options,
var  nvars,
var  FitnessFcn,
var  unused,
var  thisPopulation 
)

Crossover operator for a Genetic Algoritm.

A Simple crossover operator The arguments to the function are.

See the genetic toolbox of Matlab for more information

  • parents ??? Row vector of parents chosen by the selection function
  • options ??? options structure
  • nvars ??? Number of variables
  • FitnessFcn ??? Fitness function
  • unused ??? Placeholder not used
  • thisPopulation ??? Matrix representing the current population. The number of rows of the matrix is Population size and the number of columns is Number of variables.
function getBackend ( )
inline

Returns a user-friendly description of this factory.

Parameters
resdescription
function getBasicBatchObservables ( )
inherited

TODO.

Generate the observable objects that handle grouped data.

function getBatchObservables ( )
inlineinherited

TODO.

function getBounds ( )
inline

Returns bounds on the hyperparameters.

Return values
LBlower bound
UBupper bound
function getConstraintFcn ( )
inlineinherited

TODO.

function getCovarianceFunction ( )
inline

Returns the set of covariance functions.

Parameters
resBasisFunction classes
function getCreationFcn ( )
inlineinherited

TODO.

function getCrossoverFcn ( )
inlineinherited

TODO.

function getDimensions ( )
inlineinherited

TODO.

function getIndividualSize ( )
inlineinherited

TODO.

function getIntegerParameters ( )
inlineinherited

MATLAB GA supports handling parameters as integers, override this function in derived factories when parameters must be handled as integers.

function getLikelihoodFunction ( )
inline

Returns the set of likelihood functions.

Parameters
resBasisFunction classes
function getMeanFunction ( )
inline

Returns the set of mean functions.

Parameters
resBasisFunction classes
function getMode ( )
inlineinherited

TODO.

function getModelType ( )
inline

Returns the supported surrogate model type.

Return values
resname of the supported surrogate model class
function getMutationFcn ( )
inlineinherited

TODO.

function getObservables ( )

Defines the supported observables of this surrogate model.

Returns the observables for the Gaussian Process model.

Return values
obscell array of observables
function getParallelMode ( )
inlineinherited

TODO.

function getSamples ( )
inlineinherited

TODO.

function isCustomMode ( )
inlineinherited

TODO.

function isSliceSamplingEnabled ( )
inlineinherited

Performs elliptical slice sampling.

Parameters
hp0starting point for the sampling
likelihoodfunction handle providing the log-likelihood
function mutation ( var  parents,
var  options,
var  nvars,
var  FitnessFcn,
var  state,
var  thisScore,
var  thisPopulation 
)

Mutation operator for a Genetic Algoritm.

The arguments to the function are.

See the genetic toolbox of Matlab for more information

  • parents - Row vector of parents chosen by the selection function
  • options - Options structure
  • nvars - Number of variables
  • FitnessFcn - Fitness function
  • state - Structure containing information about the current generation.
  • thisScore - Vector of scores of the current population *thisPopulation - Matrix of individuals in the current population

The function returns mutationChildren as a matrix whose rows correspond to the children. The number of columns of the matrix is Number of variables.

function performSliceSampling ( var  hp0,
var  likelihood 
)
protectedinherited

Performs standard slice sampling.

Method implementing standard slice sampling.

Parameters
hp0starting point for the sampling
likelihoodfunction handle providing the log-likelihood
function setSamples ( var  samples,
var  values 
)
inlineinherited

TODO.

function sliceSample ( var  model)
protected

Given an individual kriging model, sample the likelihood using slice sampling.

function supportsComplexData ( )
inline

Defines whether this factory supports complex data.

Parameters
resboolean
function supportsMultipleOutputs ( )
inline

Defines whether this factory supports multiple outputs.

Parameters
resboolean
function wrapFunctions ( )
inherited

TODO.

Create anonymous function handles to wrap all operator functions to members of the given obj The genetic operator functions are memberfunctions and thus need to be called with the 'this' object (s).

To make this possible we need to wrap them in an anonymous function. And set them again as function handles.


The documentation for this class was generated from the following files: