org.freehep.math.minuit

Class MnApplication

java.lang.Object

org.freehep.math.minuit.MnApplication

Direct Known Subclasses:

MnMigrad, MnMinimize, MnScan, MnSimplex

public abstract class MnApplication
extends java.lang.Object

Base class for minimizers.

Method Summary

Modifier and Type	Method and Description
void	add(java.lang.String name, double val) add const parameter
void	add(java.lang.String name, double val, double err) add free parameter
void	add(java.lang.String name, double val, double err, double low, double up) add limited parameter
boolean	checkAnalyticalDerivatives()
MnUserCovariance	covariance()
double	error(int index)
double	error(java.lang.String name)
double	errorDef()
double[]	errors()
FCNBase	fcnbase()
void	fix(int index)
void	fix(java.lang.String name)
int	index(java.lang.String name) convert name into external number of parameter
FunctionMinimum	minimize()
FunctionMinimum	minimize(int maxfcn)
FunctionMinimum	minimize(int maxfcn, double toler) Causes minimization of the FCN and returns the result in form of a FunctionMinimum.
java.lang.String	name(int index) convert external number into name of parameter
int	numOfCalls()
MnUserParameters	parameters()
double[]	params() access to parameters and errors in column-wise representation
MnMachinePrecision	precision()
void	release(int index)
void	release(java.lang.String name)
void	removeLimits(int index)
void	removeLimits(java.lang.String name)
void	setCheckAnalyticalDerivatives(boolean check) Minuit does a check of the user gradient at the beginning, if this is not wanted the set this to "false".
void	setError(int index, double err)
void	setError(java.lang.String name, double err)
void	setErrorDef(double errorDef) errorDef() is the error definition of the function.
void	setLimits(int index, double low, double up)
void	setLimits(java.lang.String name, double low, double up)
void	setPrecision(double prec)
void	setUseAnalyticalDerivatives(boolean use) By default if the function to be minimized implements FCNGradientBase then the analytical gradient provided by the function will be used.
void	setValue(int index, double val)
void	setValue(java.lang.String name, double val)
MnUserParameterState	state()
MnStrategy	strategy()
boolean	useAnalyticalDerivaties()
double	value(int index)
double	value(java.lang.String name)
int	variableParameters()

Methods inherited from class java.lang.Object

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

Method Detail

minimize

public FunctionMinimum minimize()

minimize

public FunctionMinimum minimize(int maxfcn)

minimize

public FunctionMinimum minimize(int maxfcn,
                                double toler)

Causes minimization of the FCN and returns the result in form of a FunctionMinimum.

Parameters:

maxfcn - specifies the (approximate) maximum number of function calls after which the calculation will be stopped even if it has not yet converged.

toler - specifies the required tolerance on the function value at the minimum. The default tolerance value is 0.1, and the minimization will stop when the estimated vertical distance to the minimum (EDM) is less than 0:001*tolerance*errorDef

precision

public MnMachinePrecision precision()

state

public MnUserParameterState state()

parameters

public MnUserParameters parameters()

covariance

public MnUserCovariance covariance()

fcnbase

public FCNBase fcnbase()

strategy

public MnStrategy strategy()

numOfCalls

public int numOfCalls()

params

public double[] params()

access to parameters and errors in column-wise representation

errors

public double[] errors()

add

public void add(java.lang.String name,
                double val,
                double err)

add free parameter

add

public void add(java.lang.String name,
                double val,
                double err,
                double low,
                double up)

add limited parameter

add

public void add(java.lang.String name,
                double val)

add const parameter

fix

public void fix(int index)

release

public void release(int index)

setValue

public void setValue(int index,
                     double val)

setError

public void setError(int index,
                     double err)

setLimits

public void setLimits(int index,
                      double low,
                      double up)

removeLimits

public void removeLimits(int index)

value

public double value(int index)

error

public double error(int index)

fix

public void fix(java.lang.String name)

release

public void release(java.lang.String name)

setValue

public void setValue(java.lang.String name,
                     double val)

setError

public void setError(java.lang.String name,
                     double err)

setLimits

public void setLimits(java.lang.String name,
                      double low,
                      double up)

removeLimits

public void removeLimits(java.lang.String name)

setPrecision

public void setPrecision(double prec)

value

public double value(java.lang.String name)

error

public double error(java.lang.String name)

index

public int index(java.lang.String name)

convert name into external number of parameter

name

public java.lang.String name(int index)

convert external number into name of parameter

variableParameters

public int variableParameters()

setUseAnalyticalDerivatives

public void setUseAnalyticalDerivatives(boolean use)

By default if the function to be minimized implements FCNGradientBase then the analytical gradient provided by the function will be used. Set this to false to disable this behaviour and force numerical calculation of the gradient.

useAnalyticalDerivaties

public boolean useAnalyticalDerivaties()

setCheckAnalyticalDerivatives

public void setCheckAnalyticalDerivatives(boolean check)

Minuit does a check of the user gradient at the beginning, if this is not wanted the set this to "false".

checkAnalyticalDerivatives

public boolean checkAnalyticalDerivatives()

setErrorDef

public void setErrorDef(double errorDef)

errorDef() is the error definition of the function. E.g. is 1 if function is Chi2 and 0.5 if function is -logLikelihood. If the user wants instead the 2-sigma errors, errorDef() = 4, as Chi2(x+n*sigma) = Chi2(x) + n*n.

errorDef

public double errorDef()