de.jstacs.algorithms.optimization.termination
Class SmallDifferenceOfFunctionEvaluationsCondition

java.lang.Object
  de.jstacs.algorithms.optimization.termination.AbstractTerminationCondition
      de.jstacs.algorithms.optimization.termination.SmallDifferenceOfFunctionEvaluationsCondition

All Implemented Interfaces:

TerminationCondition, InstantiableFromParameterSet, Storable, Cloneable

public class SmallDifferenceOfFunctionEvaluationsCondition
extends AbstractTerminationCondition

This class implements a TerminationCondition that stops an optimization if the difference of the current and the last function evaluations will be small, i.e., .

Author:

Jens Keilwagen

Nested Class Summary

static class

SmallDifferenceOfFunctionEvaluationsCondition.SmallDifferenceOfFunctionEvaluationsConditionParameterSet This class implements the parameter set for a SmallDifferenceOfFunctionEvaluationsCondition.

Nested classes/interfaces inherited from class de.jstacs.algorithms.optimization.termination.AbstractTerminationCondition

AbstractTerminationCondition.AbstractTerminationConditionParameterSet

Field Summary

Fields inherited from class de.jstacs.algorithms.optimization.termination.AbstractTerminationCondition

parameter

Constructor Summary

SmallDifferenceOfFunctionEvaluationsCondition(double epsilon)
This constructor creates an instance that stops the optimization if the difference of the current and the last function evaluations is smaller than epsilon.

SmallDifferenceOfFunctionEvaluationsCondition(SmallDifferenceOfFunctionEvaluationsCondition.SmallDifferenceOfFunctionEvaluationsConditionParameterSet parameter)
This is the main constructor creating an instance from a given parameter set.

SmallDifferenceOfFunctionEvaluationsCondition(StringBuffer xml)
The standard constructor for the interface Storable.

Method Summary

boolean	doNextIteration(int iteration, double f_last, double f_current, double[] gradient, double[] direction, double alpha, Time t) This method allows to decide whether to do another iteration in an optimization or not.
protected String	getXmlTag() This method returns the xml tag that is used in the method AbstractTerminationCondition.toXML() and in the constructor AbstractTerminationCondition.AbstractTerminationCondition(StringBuffer).
boolean	isSimple() This method returns false if the TerminationCondition uses either the gradient or the direction for the decision, otherwise it returns true.
protected void	set() This method sets internal member variables from AbstractTerminationCondition.parameter.

Methods inherited from class de.jstacs.algorithms.optimization.termination.AbstractTerminationCondition

clone, getCurrentParameterSet, toXML

Methods inherited from class java.lang.Object

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

Constructor Detail

SmallDifferenceOfFunctionEvaluationsCondition

public SmallDifferenceOfFunctionEvaluationsCondition(double epsilon)
                                              throws Exception

This constructor creates an instance that stops the optimization if the difference of the current and the last function evaluations is smaller than epsilon.

Parameters:

epsilon - the threshold for stopping, has to be non-negative

Throws:

Exception - if the parameter can not be set properly

SmallDifferenceOfFunctionEvaluationsCondition

public SmallDifferenceOfFunctionEvaluationsCondition(SmallDifferenceOfFunctionEvaluationsCondition.SmallDifferenceOfFunctionEvaluationsConditionParameterSet parameter)
                                              throws CloneNotSupportedException

This is the main constructor creating an instance from a given parameter set.

Parameters:

parameter - the set of parameters

Throws:

CloneNotSupportedException - if parameter can not be cloned properly.

SmallDifferenceOfFunctionEvaluationsCondition

public SmallDifferenceOfFunctionEvaluationsCondition(StringBuffer xml)
                                              throws NonParsableException

The standard constructor for the interface Storable. Constructs a SmallDifferenceOfFunctionEvaluationsCondition out of an XML representation.

Parameters:

xml - the XML representation as StringBuffer

Throws:

NonParsableException - if the SmallDifferenceOfFunctionEvaluationsCondition could not be reconstructed out of the StringBuffer xml

Method Detail

set

protected void set()

Description copied from class: AbstractTerminationCondition

This method sets internal member variables from AbstractTerminationCondition.parameter. It is used in the constructors.

Specified by:

set in class AbstractTerminationCondition

getXmlTag

protected String getXmlTag()

Description copied from class: AbstractTerminationCondition

This method returns the xml tag that is used in the method AbstractTerminationCondition.toXML() and in the constructor AbstractTerminationCondition.AbstractTerminationCondition(StringBuffer).

Specified by:

getXmlTag in class AbstractTerminationCondition

Returns:

the xml tag of the instance

doNextIteration

public boolean doNextIteration(int iteration,
                               double f_last,
                               double f_current,
                               double[] gradient,
                               double[] direction,
                               double alpha,
                               Time t)

Description copied from interface: TerminationCondition

This method allows to decide whether to do another iteration in an optimization or not. If it returns true it is recommended to do another iteration, otherwise (the method returns false and) it is recommended to stop the algorithm.

Parameters:

iteration - the number of performed iterations

f_last - last value of the function

f_current - current value of the function

gradient - the gradient of the function

direction - the last direction of the optimization

alpha - the last step size

t - a time object measuring the time that has been elapsed in the optimization

Returns:

true if another iteration should be done

isSimple

public boolean isSimple()

Description copied from interface: TerminationCondition

This method returns false if the TerminationCondition uses either the gradient or the direction for the decision, otherwise it returns true.

Returns:

false for gradient or direction based TerminationConditions