org.apache.commons.math.optimization.direct
Class MultiDirectional

java.lang.Object
  org.apache.commons.math.optimization.direct.DirectSearchOptimizer
      org.apache.commons.math.optimization.direct.MultiDirectional

All Implemented Interfaces:

MultivariateRealOptimizer

public class MultiDirectional
extends DirectSearchOptimizer

This class implements the multi-directional direct search method.

Since:

1.2

Version:

$Revision: 1070725 $ $Date: 2011-02-15 02:31:12 +0100 (mar. 15 févr. 2011) $

See Also:

NelderMead

Field Summary

private double	gamma Contraction coefficient.
private double	khi Expansion coefficient.

Fields inherited from class org.apache.commons.math.optimization.direct.DirectSearchOptimizer

simplex

Constructor Summary

MultiDirectional()
Build a multi-directional optimizer with default coefficients.

MultiDirectional(double khi, double gamma)
Build a multi-directional optimizer with specified coefficients.

Method Summary

private RealPointValuePair	evaluateNewSimplex(RealPointValuePair[] original, double coeff, java.util.Comparator<RealPointValuePair> comparator) Compute and evaluate a new simplex.
protected void	iterateSimplex(java.util.Comparator<RealPointValuePair> comparator) Compute the next simplex of the algorithm.

Methods inherited from class org.apache.commons.math.optimization.direct.DirectSearchOptimizer

evaluate, evaluateSimplex, getConvergenceChecker, getEvaluations, getIterations, getMaxEvaluations, getMaxIterations, incrementIterationsCounter, optimize, replaceWorstPoint, setConvergenceChecker, setMaxEvaluations, setMaxIterations, setStartConfiguration, setStartConfiguration

Methods inherited from class java.lang.Object

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

Field Detail

khi

private final double khi

Expansion coefficient.

gamma

private final double gamma

Contraction coefficient.

Constructor Detail

MultiDirectional

public MultiDirectional()

Build a multi-directional optimizer with default coefficients.

The default values are 2.0 for khi and 0.5 for gamma.

MultiDirectional

public MultiDirectional(double khi,
                        double gamma)

Build a multi-directional optimizer with specified coefficients.

Parameters:

khi - expansion coefficient

gamma - contraction coefficient

Method Detail

iterateSimplex

protected void iterateSimplex(java.util.Comparator<RealPointValuePair> comparator)
                       throws FunctionEvaluationException,
                              OptimizationException,
                              java.lang.IllegalArgumentException

Compute the next simplex of the algorithm.

Specified by:

iterateSimplex in class DirectSearchOptimizer

Parameters:

comparator - comparator to use to sort simplex vertices from best to worst

Throws:

FunctionEvaluationException - if the function cannot be evaluated at some point

OptimizationException - if the algorithm fails to converge

java.lang.IllegalArgumentException - if the start point dimension is wrong

evaluateNewSimplex

private RealPointValuePair evaluateNewSimplex(RealPointValuePair[] original,
                                              double coeff,
                                              java.util.Comparator<RealPointValuePair> comparator)
                                       throws FunctionEvaluationException,
                                              OptimizationException

Compute and evaluate a new simplex.

Parameters:

original - original simplex (to be preserved)

coeff - linear coefficient

comparator - comparator to use to sort simplex vertices from best to poorest

Returns:

best point in the transformed simplex

Throws:

FunctionEvaluationException - if the function cannot be evaluated at some point

OptimizationException - if the maximal number of evaluations is exceeded