jebl.math
Class OrthogonalSearch

java.lang.Object
  jebl.math.MultivariateMinimum
      jebl.math.OrthogonalSearch

public class OrthogonalSearch
extends MultivariateMinimum

minimization of a real-valued function of several variables without using derivatives, using the simple strategy of optimizing variables one by one.

Author:

Korbinian Strimmer, Matthew Goode

Nested Class Summary

protected static interface

OrthogonalSearch.RoundOptimiser

Nested classes/interfaces inherited from class jebl.math.MultivariateMinimum

MultivariateMinimum.Factory

Field Summary

Fields inherited from class jebl.math.MultivariateMinimum

maxFun, numFun, numFuncStops

Constructor Summary

OrthogonalSearch()
Initialization

OrthogonalSearch(boolean shuffle)
Initialization

OrthogonalSearch(OrderEnumerator.OEFactory orderingFactory)
Initialization

Method Summary

protected void	debug(Object output)
static MultivariateMinimum.Factory	generateFactory(boolean shuffle) Generate a MultivariateMinimum.Factory for an OrthogonalSearch
protected OrthogonalSearch.RoundOptimiser	generateOrthogonalRoundOptimiser(MultivariateFunction mf)
protected UnivariateMinimum	generateUnivariateMinimum()
protected boolean	isDebug()
protected boolean	isFrequentMonitoring()
protected boolean	isIgnoreNonMinimalUnivariateMinimisations() Should we ignore new minisations that are not as minimal as the current one?
protected boolean	isUseCurrentInUnivariateMinimisation()
void	optimize(MultivariateFunction f, double[] xvec, double tolfx, double tolx) The actual optimization routine (needs to be implemented in a subclass of MultivariateMinimum).
void	optimize(MultivariateFunction f, double[] xvec, double tolfx, double tolx, MinimiserMonitor monitor) The actual optimization routine It finds a minimum close to vector x when the absolute tolerance for each parameter is specified.
void	setIgnoreNonMinimalUnivariateMinimisations(boolean value) Should we ignore new minisations that are not as minimal as the current one?
void	setUseCurrentInUnivariateMinimisation(boolean value)

Methods inherited from class jebl.math.MultivariateMinimum

copy, findMinimum, findMinimum, findMinimum, stopCondition

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

OrthogonalSearch

public OrthogonalSearch()

Initialization

OrthogonalSearch

public OrthogonalSearch(boolean shuffle)

Initialization

Parameters:

shuffle - If true uses shuffling, else uses ascending order, when choosing next parameter to optimse (true means equivalent to old StochasticOSearch)

OrthogonalSearch

public OrthogonalSearch(OrderEnumerator.OEFactory orderingFactory)

Initialization

Method Detail

setUseCurrentInUnivariateMinimisation

public void setUseCurrentInUnivariateMinimisation(boolean value)

setIgnoreNonMinimalUnivariateMinimisations

public void setIgnoreNonMinimalUnivariateMinimisations(boolean value)

Should we ignore new minisations that are not as minimal as the current one?

optimize

public void optimize(MultivariateFunction f,
                     double[] xvec,
                     double tolfx,
                     double tolx)

Description copied from class: MultivariateMinimum

The actual optimization routine (needs to be implemented in a subclass of MultivariateMinimum). It finds a minimum close to vector x when the absolute tolerance for each parameter is specified.

Specified by:

optimize in class MultivariateMinimum

Parameters:

f - multivariate function

xvec - initial guesses for the minimum (contains the location of the minimum on return)

tolfx - absolute tolerance of function value

tolx - absolute tolerance of each parameter

optimize

public void optimize(MultivariateFunction f,
                     double[] xvec,
                     double tolfx,
                     double tolx,
                     MinimiserMonitor monitor)

Description copied from class: MultivariateMinimum

The actual optimization routine It finds a minimum close to vector x when the absolute tolerance for each parameter is specified.

Overrides:

optimize in class MultivariateMinimum

Parameters:

f - multivariate function

xvec - initial guesses for the minimum (contains the location of the minimum on return)

tolfx - absolute tolerance of function value

tolx - absolute tolerance of each parameter

monitor - A monitor object that receives information about the minimising process (for display purposes) note: The default implementation just calls the optimize function with out the Monitor!

generateFactory

public static final MultivariateMinimum.Factory generateFactory(boolean shuffle)

Generate a MultivariateMinimum.Factory for an OrthogonalSearch

Parameters:

shuffle - if true shuffles order for each round (see OrthogonalSearch constructors)

generateUnivariateMinimum

protected UnivariateMinimum generateUnivariateMinimum()

isFrequentMonitoring

protected boolean isFrequentMonitoring()

generateOrthogonalRoundOptimiser

protected OrthogonalSearch.RoundOptimiser generateOrthogonalRoundOptimiser(MultivariateFunction mf)

isUseCurrentInUnivariateMinimisation

protected final boolean isUseCurrentInUnivariateMinimisation()

isIgnoreNonMinimalUnivariateMinimisations

protected final boolean isIgnoreNonMinimalUnivariateMinimisations()

Should we ignore new minisations that are not as minimal as the current one?

debug

protected void debug(Object output)

isDebug

protected boolean isDebug()