Description of the class TOptimizer. More...

#include <TOptimizer.h>

Inheritance diagram for URANIE::Optimizer::TOptimizer:

Collaboration diagram for URANIE::Optimizer::TOptimizer:

Public Types
enum	EOptimType { kMinimizeCode, kMinimizeFunction, kSumOfSquare }

enum	EOptimMethod { kMigrad, kSimplex, kCombined, kScan }

Public Member Functions
Constructor and Destructor
	TOptimizer (URANIE::DataServer::TDataServer tds, URANIE::Launcher::TCode code)
	Constructor with a dataserver. More...

	TOptimizer (URANIE::DataServer::TDataServer tds, void(fcn)(Double_t , Double_t ), TString sinput, TString soutput)
	Constructor by a TDataServer, a function and list of input and output (optional) More...

	TOptimizer (URANIE::DataServer::TDataServer tds, const char fcn, TString sinput="", TString soutput="")
	Constructor by a TDataServer, a function name and list of input and output (optional) More...

virtual	~TOptimizer ()
	Default destructor. More...

Set/Get parameters
void	setTolerance (Double_t dtol)
	Set the tolerance parameter. More...

Double_t	getTolerance ()
	Get the tolerance parameter. More...

void	setMaxIterations (Int_t nmax)
	Set the maximum number of iterations. More...

Int_t	getMaxIterations ()
	Get the maximum number of iterations. More...

void	setMaxFunctionCalls (Int_t nmax)
	Set the maximum number of evaluations. More...

Int_t	getMaxFunctionCalls ()
	Get the maximum numbers of evaluations. More...

Set/Get the method of Optimization
These classe is a wrapper with the TMinuit library interfaced in ROOT which contains several algorithm of optimization : kMigrad : the Migrad algorithm; it is a quasi Newton algorithm. It is the default algorithm selected in Uranie. kSimplex : the Simplex algorithm kCombined : kScan :
void	setMethod (EOptimMethod method=kMigrad)
	Select the algorithm of optimization. More...

EOptimMethod	getMethod ()
	Return the algorithm of optimization. More...

EOptimType	getOptimType ()
	Return the type of optimization. More...

void	setPrintLevel (int level)
	Set the print Level of Debug between \( [1, 5]\). More...

int	getPrintLevel () const
	Get the print Level of Debug between \( [1, 5]\). More...

Manipulate the code
virtual void	init (Option_t *option="")
	Init phase of the code. More...

virtual void	optimize (Option_t *option="")
	Runs phase of the code (TMinuit2) More...

virtual void	terminate (Option_t *option="")
	Terminates phase of the code. More...

virtual void	clean (Option_t *option="")
	Cleans the working directory of the code. More...

Manipulate the Objectives
In the case of calibration of code, we can define one or several "Objectives" to specify the criteria to minimize. The criteria to minimize is a sum of weighted objectives: \[ C = \sum_{i=1}^{nO}{\omega_i Obj_{i}} \] By default, all added objectives are "active" and the weight \( \omega = 1.0\).
void	addObjective (TString name, URANIE::DataServer::TDataServer tds, TString ystar, URANIE::Launcher::TOutputFile outfile, TString yhat, Double_t weight=1.0)
	Add the objective in the TOptimizer object : Sum of square of the error. More...

void	addObjective (TString name, URANIE::DataServer::TDataServer tds, TString ystar, TString sigma, URANIE::Launcher::TOutputFile outfile, TString yhat, Double_t alpha=1.0)
	Add the objective in the TOptimizer object : Sum of square of the weighted error. More...

void	unactiveObjective (TString name)
	Unactive the objective given by these name. More...

void	activeObjective (TString name)
	Active the objective given by these name. More...

void	invertObjectives ()
	Invert the state of all the objectives. More...

vector< URANIE::Optimizer::TObjective *>	getObjectives ()
	get the vector containing the objectives More...

void	addOutputVariable (TString soutput)
	Add input and output variables. More...

void	selectCost (TString scost)
	Select an cost in the Output string. More...

TString	getSelectedCost ()
	Get the selected Select an cost in the Output string. More...

Manipulate the parameters
At any time, we can change the state of the parameters; it can be fix to the default value or it can be variable in these range.
void	fixParameter (TString name)
	Fix the parameter given by these name to the default value. More...

void	unfixParameter (TString name)
	Unfix the parameter given by these name. More...

void	invertParameters ()
	Invert the state of all the parameters. More...

Printing Log
void	printState (Option_t *option="")
	Print the state of the algorithm of optimization. More...

virtual void	printLog (Option_t *option="")

Private Attributes
Bool_t	_bsave
	Saves all the computation files in each directory. More...

Bool_t	_bclean
	Cleans the Working Directory before running the code. More...

Int_t	_nPrintLevel
	The level of Debg between \( [1, 5]\) without test. More...

URANIE::DataServer::TDataServer *	_tds

Int_t	_nMaxIterations
	Pointer vers un TDS. More...

Int_t	_nMaxFunctionCalls
	The max number to call the code/function. More...

Double_t	_dTolerance
	The tolerance parameter. More...

vector< TObjective * >	_objectives
	vector of objectives More...

TMethodCall *	_fMethodCall
	Pointer to MethodCall in case of interpreted function. More...

void(*	_fMethod )(Double_t , Double_t )

TString	_sFunctionName
	A pointer for the analytical function. More...

EOptimType	_optimProblem
	Function name. More...

EOptimMethod	_optimMethod
	The method of optimisation. More...

TString	_sInput
	The string of input attributes. More...

TString	_sOutput
	The string of output attributes (the number must be equal to 1) More...

TString	_sCost
	The name od the selected cost (the number must be equal to 1) More...

Detailed Description

Description of the class TOptimizer.

To be written by the developper.

Member Enumeration Documentation

◆ EOptimMethod

enum URANIE::Optimizer::TOptimizer::EOptimMethod

Enumerator
kMigrad
kSimplex
kCombined
kScan

◆ EOptimType

enum URANIE::Optimizer::TOptimizer::EOptimType

Enumerator
kMinimizeCode
kMinimizeFunction
kSumOfSquare

Constructor & Destructor Documentation

◆ TOptimizer() [1/3]

URANIE::Optimizer::TOptimizer::TOptimizer	(	URANIE::DataServer::TDataServer *	tds,
		URANIE::Launcher::TCode *	code
	)

Constructor with a dataserver.

Referenced by ClassImp().

◆ TOptimizer() [2/3]

URANIE::Optimizer::TOptimizer::TOptimizer	(	URANIE::DataServer::TDataServer *	tds,
		void()(Double_t , Double_t *)	fcn,
		TString	sinput,
		TString	soutput
	)

Constructor by a TDataServer, a function and list of input and output (optional)

◆ TOptimizer() [3/3]

URANIE::Optimizer::TOptimizer::TOptimizer	(	URANIE::DataServer::TDataServer *	tds,
		const char *	fcn,
		TString	sinput = `""`,
		TString	soutput = `""`
	)

Constructor by a TDataServer, a function name and list of input and output (optional)

◆ ~TOptimizer()

virtual URANIE::Optimizer::TOptimizer::~TOptimizer ( )

virtual

Default destructor.

Referenced by ClassImp().

Member Function Documentation

◆ activeObjective()

void URANIE::Optimizer::TOptimizer::activeObjective ( TString name )

Active the objective given by these name.

Referenced by ClassImp().

◆ addObjective() [1/2]

void URANIE::Optimizer::TOptimizer::addObjective	(	TString	name,
		URANIE::DataServer::TDataServer *	tds,
		TString	ystar,
		URANIE::Launcher::TOutputFile *	outfile,
		TString	yhat,
		Double_t	weight = `1.0`
	)

Add the objective in the TOptimizer object : Sum of square of the error.

The objective is computed with formula

\[ obj = \alpha \sum_{i=1}^{n}{(y^{\star}_{i} - \hat{y}_{i})^{2}} \]

Parameters

name	(TString) the name of the TObjective
tds	(TDataServer *) The pointer of the TDataServer object which contains the attribute \( y^{\star}\)
ystar	(TString) the name of the \( y^{\star}\) attribute
outfile	(TOutputFile *) The pointer of the TOutputFile object which contains the attribute \( \hat{y}\)
yhat	(TString) the name of the \( \hat{y}\) attribute
weight	(Double_t) The weight of the objective [1.0]

Referenced by ClassImp().

◆ addObjective() [2/2]

void URANIE::Optimizer::TOptimizer::addObjective	(	TString	name,
		URANIE::DataServer::TDataServer *	tds,
		TString	ystar,
		TString	sigma,
		URANIE::Launcher::TOutputFile *	outfile,
		TString	yhat,
		Double_t	alpha = `1.0`
	)

Add the objective in the TOptimizer object : Sum of square of the weighted error.

The objective is computed with formula

\[ obj = \alpha \sum_{i=1}^{n}{\frac{(y^{\star}_{i} - \hat{y}_{i})^{2}}{\sigma_{i}^{2}}} \]

Parameters

name	(TString) the name of the TObjective
tds	(TDataServer *) The pointer of the TDataServer object which contains the attribute \( y^{\star}\)
ystar	(TString) the name of the \( y^{\star}\) attribute
sigma	(TString) the name of the weight \( \sigma\) attribute
outfile	(TOutputFile *) The pointer of the TOutputFile object which contains the attribute \( \hat{y}\)
yhat	(TString) the name of the \( \hat{y}\) attribute
alpha	(Double_t) The weight of the objective [1.0]

◆ addOutputVariable()

void URANIE::Optimizer::TOptimizer::addOutputVariable ( TString soutput )

Add input and output variables.

Add output attributes defined from other attributes

Parameters

soutput (TString): list of output variable to be created

Referenced by ClassImp().

◆ clean()

virtual void URANIE::Optimizer::TOptimizer::clean ( Option_t * option = "" )

virtual

Cleans the working directory of the code.

Referenced by ClassImp().

◆ fixParameter()

void URANIE::Optimizer::TOptimizer::fixParameter ( TString name )

Fix the parameter given by these name to the default value.

Referenced by ClassImp().

◆ getMaxFunctionCalls()

Int_t URANIE::Optimizer::TOptimizer::getMaxFunctionCalls ( )

inline

Get the maximum numbers of evaluations.

References _nMaxFunctionCalls.

◆ getMaxIterations()

Int_t URANIE::Optimizer::TOptimizer::getMaxIterations ( )

inline

Get the maximum number of iterations.

References _nMaxIterations.

◆ getMethod()

EOptimMethod URANIE::Optimizer::TOptimizer::getMethod ( )

inline

Return the algorithm of optimization.

References _optimMethod.

◆ getObjectives()

vector< URANIE::Optimizer::TObjective * > URANIE::Optimizer::TOptimizer::getObjectives ( )

inline

get the vector containing the objectives

References _objectives.

◆ getOptimType()

EOptimType URANIE::Optimizer::TOptimizer::getOptimType ( )

inline

Return the type of optimization.

References _optimProblem.

◆ getPrintLevel()

int URANIE::Optimizer::TOptimizer::getPrintLevel ( ) const

inline

Get the print Level of Debug between \( [1, 5]\).

References _nPrintLevel.

◆ getSelectedCost()

TString URANIE::Optimizer::TOptimizer::getSelectedCost ( )

inline

Get the selected Select an cost in the Output string.

References _sCost.

◆ getTolerance()

Double_t URANIE::Optimizer::TOptimizer::getTolerance ( )

inline

Get the tolerance parameter.

References _dTolerance.

◆ init()

virtual void URANIE::Optimizer::TOptimizer::init ( Option_t * option = "" )

virtual

Init phase of the code.

Referenced by ClassImp().

◆ invertObjectives()

void URANIE::Optimizer::TOptimizer::invertObjectives ( )

Invert the state of all the objectives.

Referenced by ClassImp().

◆ invertParameters()

void URANIE::Optimizer::TOptimizer::invertParameters ( )

Invert the state of all the parameters.

Then, fixed parameters became "variable" attribute and vice versa.

Referenced by ClassImp().

◆ optimize()

virtual void URANIE::Optimizer::TOptimizer::optimize ( Option_t * option = "" )

virtual

Runs phase of the code (TMinuit2)

Referenced by ClassImp().

◆ printLog()

virtual void URANIE::Optimizer::TOptimizer::printLog ( Option_t * option = "" )

virtual

Referenced by ClassImp().

◆ printState()

void URANIE::Optimizer::TOptimizer::printState ( Option_t * option = "" )

Print the state of the algorithm of optimization.

Referenced by ClassImp().

◆ selectCost()

void URANIE::Optimizer::TOptimizer::selectCost ( TString scost )

Select an cost in the Output string.

Referenced by ClassImp().

◆ setMaxFunctionCalls()

void URANIE::Optimizer::TOptimizer::setMaxFunctionCalls ( Int_t nmax )

inline

Set the maximum number of evaluations.

References _nMaxFunctionCalls.

◆ setMaxIterations()

void URANIE::Optimizer::TOptimizer::setMaxIterations ( Int_t nmax )

inline

Set the maximum number of iterations.

References _nMaxIterations.

◆ setMethod()

void URANIE::Optimizer::TOptimizer::setMethod ( EOptimMethod method = kMigrad )

inline

Select the algorithm of optimization.

◆ setPrintLevel()

void URANIE::Optimizer::TOptimizer::setPrintLevel ( int level )

inline

Set the print Level of Debug between \( [1, 5]\).

References _nPrintLevel.

◆ setTolerance()

void URANIE::Optimizer::TOptimizer::setTolerance ( Double_t dtol )

inline

Set the tolerance parameter.

◆ terminate()

virtual void URANIE::Optimizer::TOptimizer::terminate ( Option_t * option = "" )

virtual

Terminates phase of the code.

Referenced by ClassImp().

◆ unactiveObjective()

void URANIE::Optimizer::TOptimizer::unactiveObjective ( TString name )

Unactive the objective given by these name.

Referenced by ClassImp().

◆ unfixParameter()

void URANIE::Optimizer::TOptimizer::unfixParameter ( TString name )

Unfix the parameter given by these name.

These parameter became a "variable" attribute

Referenced by ClassImp().

Member Data Documentation

◆ _bclean

Bool_t URANIE::Optimizer::TOptimizer::_bclean

private

Cleans the Working Directory before running the code.

Referenced by ClassImp().

◆ _bsave

Bool_t URANIE::Optimizer::TOptimizer::_bsave

private

Saves all the computation files in each directory.

Referenced by ClassImp().

◆ _dTolerance

Double_t URANIE::Optimizer::TOptimizer::_dTolerance

private

The tolerance parameter.

Referenced by ClassImp(), and getTolerance().

◆ _fMethod

void(* URANIE::Optimizer::TOptimizer::_fMethod) (Double_t *, Double_t *)

private

Referenced by ClassImp().

◆ _fMethodCall

TMethodCall* URANIE::Optimizer::TOptimizer::_fMethodCall

private

Pointer to MethodCall in case of interpreted function.

◆ _nMaxFunctionCalls

Int_t URANIE::Optimizer::TOptimizer::_nMaxFunctionCalls

private

The max number to call the code/function.

Referenced by ClassImp(), getMaxFunctionCalls(), and setMaxFunctionCalls().

◆ _nMaxIterations

Int_t URANIE::Optimizer::TOptimizer::_nMaxIterations

private

Pointer vers un TDS.

The max number of iterations

Referenced by ClassImp(), getMaxIterations(), and setMaxIterations().

◆ _nPrintLevel

Int_t URANIE::Optimizer::TOptimizer::_nPrintLevel

private

The level of Debg between \( [1, 5]\) without test.

Referenced by ClassImp(), getPrintLevel(), and setPrintLevel().

◆ _objectives

vector<TObjective *> URANIE::Optimizer::TOptimizer::_objectives

private

vector of objectives

Referenced by ClassImp(), and getObjectives().

◆ _optimMethod

EOptimMethod URANIE::Optimizer::TOptimizer::_optimMethod

private

The method of optimisation.

Referenced by ClassImp(), and getMethod().

◆ _optimProblem

EOptimType URANIE::Optimizer::TOptimizer::_optimProblem

private

Function name.

The problem of optimisation

Referenced by ClassImp(), and getOptimType().

◆ _sCost

TString URANIE::Optimizer::TOptimizer::_sCost

private

The name od the selected cost (the number must be equal to 1)

Referenced by ClassImp(), and getSelectedCost().

◆ _sFunctionName

TString URANIE::Optimizer::TOptimizer::_sFunctionName

private

A pointer for the analytical function.

Referenced by ClassImp().

◆ _sInput

TString URANIE::Optimizer::TOptimizer::_sInput

private

The string of input attributes.

Referenced by ClassImp().

◆ _sOutput

TString URANIE::Optimizer::TOptimizer::_sOutput

private

The string of output attributes (the number must be equal to 1)

Referenced by ClassImp().

◆ _tds

URANIE::DataServer::TDataServer* URANIE::Optimizer::TOptimizer::_tds

private

Referenced by ClassImp().

Documentation / Manuel développeur

Modules disponibles

Public Types

Public Member Functions

Private Attributes

Detailed Description

Member Enumeration Documentation

◆ EOptimMethod

◆ EOptimType

Constructor & Destructor Documentation

◆ TOptimizer() [1/3]

◆ TOptimizer() [2/3]

◆ TOptimizer() [3/3]

◆ ~TOptimizer()

Member Function Documentation

◆ activeObjective()

◆ addObjective() [1/2]

◆ addObjective() [2/2]

◆ addOutputVariable()

◆ clean()

◆ fixParameter()

◆ getMaxFunctionCalls()

◆ getMaxIterations()

◆ getMethod()

◆ getObjectives()

◆ getOptimType()

◆ getPrintLevel()

◆ getSelectedCost()

◆ getTolerance()

◆ init()

◆ invertObjectives()

◆ invertParameters()

◆ optimize()

◆ printLog()

◆ printState()

◆ selectCost()

◆ setMaxFunctionCalls()

◆ setMaxIterations()

◆ setMethod()

◆ setPrintLevel()

◆ setTolerance()

◆ terminate()

◆ unactiveObjective()

◆ unfixParameter()

Member Data Documentation

◆ _bclean

◆ _bsave

◆ _dTolerance

◆ _fMethod

◆ _fMethodCall

◆ _nMaxFunctionCalls

◆ _nMaxIterations

◆ _nPrintLevel

◆ _objectives

◆ _optimMethod

◆ _optimProblem

◆ _sCost

◆ _sFunctionName

◆ _sInput

◆ _sOutput

◆ _tds