pal.math

Class MultivariateMinimum

java.lang.Object

pal.math.MultivariateMinimum

Direct Known Subclasses:

ConjugateDirectionSearch, ConjugateGradientSearch, DifferentialEvolution, GeneralizedDEOptimizer, OrthogonalSearch, StochasticOSearch

public abstract class MultivariateMinimum
extends Object

abstract base class for minimisation of a multivariate function

Author:

Korbinian Strimmer

Field Summary

Fields

Modifier and Type	Field and Description
int	maxFun maxFun is the maximum number of calls to fun allowed.
int	numFun total number of function evaluations necessary
int	numFuncStops numFuncStops is the number of consecutive positive evaluations of the stop criterion based on function evaluation necessary to cause the abortion of the optimization (default is 4)

Constructor Summary

Constructors

Constructor and Description
MultivariateMinimum()

Method Summary

Modifier and Type	Method and Description
void	copy(double[] target, double[] source) Copy source vector into target vector
double	findMinimum(MultivariateFunction f, double[] xvec) Find minimum close to vector x
double	findMinimum(MultivariateFunction f, double[] xvec, int fxFracDigits, int xFracDigits) Find minimum close to vector x (desired fractional digits for each parameter is specified)
abstract void	optimize(MultivariateFunction f, double[] xvec, double tolfx, double tolx) The actual optimization routine (needs to be implemented in a subclass of MultivariateMinimum).
boolean	stopCondition(double fx, double[] x, double tolfx, double tolx, boolean firstCall) Checks whether optimization should stop

Methods inherited from class java.lang.Object

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

Field Detail

numFun

public int numFun

total number of function evaluations necessary

maxFun

public int maxFun

maxFun is the maximum number of calls to fun allowed. the default value of 0 indicates no limit on the number of calls.

numFuncStops

public int numFuncStops

numFuncStops is the number of consecutive positive evaluations of the stop criterion based on function evaluation necessary to cause the abortion of the optimization (default is 4)

Constructor Detail

MultivariateMinimum

public MultivariateMinimum()

Method Detail

findMinimum

public double findMinimum(MultivariateFunction f,
                          double[] xvec)

Find minimum close to vector x

Parameters:

f - multivariate function

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

Returns:

minimal function value

findMinimum

public double findMinimum(MultivariateFunction f,
                          double[] xvec,
                          int fxFracDigits,
                          int xFracDigits)

Find minimum close to vector x (desired fractional digits for each parameter is specified)

Parameters:

f - multivariate function

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

fxFracDigits - desired fractional digits in the function value

xFracDigits - desired fractional digits in parameters x

Returns:

minimal function value

optimize

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

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.

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

stopCondition

public boolean stopCondition(double fx,
                             double[] x,
                             double tolfx,
                             double tolx,
                             boolean firstCall)

Checks whether optimization should stop

Parameters:

fx - current function value

x - current values of function parameters

tolfx - absolute tolerance of function value

tolx - absolute tolerance of each parameter

firstCall - needs to be set to true when this routine is first called otherwise it should be set to false

Returns:

true if either x and its previous value are sufficiently similar or if fx and its previous values are sufficiently similar (test on function value has to be succesful numFuncStops consecutive times)

copy

public void copy(double[] target,
                 double[] source)

Copy source vector into target vector

Parameters:

target - parameter array

source - parameter array