Logo ROOT  
Reference Guide
MinimTransformFunction.cxx
Go to the documentation of this file.
1// @(#)root/mathmore:$Id$
2// Author: L. Moneta June 2009
3
4/**********************************************************************
5 * *
6 * Copyright (c) 2006 LCG ROOT Math Team, CERN/PH-SFT *
7 * *
8 * *
9 **********************************************************************/
10
11// Implementation file for class MinimTransformFunction
12
14#include "Math/IFunctionfwd.h"
15
16//#include <iostream>
17#include <cmath>
18#include <cassert>
19
20namespace ROOT {
21
22 namespace Math {
23
24MinimTransformFunction::MinimTransformFunction ( const IMultiGradFunction * f, const std::vector<EMinimVariableType> & types,
25 const std::vector<double> & values,
26 const std::map<unsigned int, std::pair<double, double> > & bounds) :
27 fX( values ),
28 fFunc(f)
29{
30 // constructor of the class from a pointer to the function (which is managed)
31 // vector specifying the variable types (free, bounded or fixed, defined in enum EMinimVariableTypes )
32 // variable values (used for the fixed ones) and a map with the bounds (for the bounded variables)
33
34 unsigned int ntot = NTot(); // NTot is fFunc->NDim()
35 assert ( types.size() == ntot );
36 fVariables.reserve(ntot);
37 fIndex.reserve(ntot);
38 for (unsigned int i = 0; i < ntot; ++i ) {
39 if (types[i] == kFix )
40 fVariables.push_back( MinimTransformVariable( values[i]) );
41 else {
42 fIndex.push_back(i);
43
44 if ( types[i] == kDefault)
46 else {
47 std::map<unsigned int, std::pair<double,double> >::const_iterator itr = bounds.find(i);
48 assert ( itr != bounds.end() );
49 double low = itr->second.first;
50 double up = itr->second.second;
51 if (types[i] == kBounds )
53 else if (types[i] == kLowBound)
55 else if (types[i] == kUpBound)
57 }
58 }
59 }
60}
61
62
63void MinimTransformFunction::Transformation( const double * x, double * xext) const {
64 // transform from internal to external
65
66 unsigned int nfree = fIndex.size();
67
68// std::cout << "Transform: internal ";
69// for (int i = 0; i < nfree; ++i) std::cout << x[i] << " ";
70// std::cout << "\t\t";
71
72 for (unsigned int i = 0; i < nfree; ++i ) {
73 unsigned int extIndex = fIndex[i];
74 const MinimTransformVariable & var = fVariables[ extIndex ];
75 if (var.IsLimited() )
76 xext[ extIndex ] = var.InternalToExternal( x[i] );
77 else
78 xext[ extIndex ] = x[i];
79 }
80
81// std::cout << "Transform: external ";
82// for (int i = 0; i < fX.size(); ++i) std::cout << fX[i] << " ";
83// std::cout << "\n";
84
85}
86
87void MinimTransformFunction::InvTransformation(const double * xExt, double * xInt) const {
88 // inverse function transformation (external -> internal)
89 for (unsigned int i = 0; i < NDim(); ++i ) {
90 unsigned int extIndex = fIndex[i];
91 const MinimTransformVariable & var = fVariables[ extIndex ];
92 assert ( !var.IsFixed() );
93 if (var.IsLimited() )
94 xInt[ i ] = var.ExternalToInternal( xExt[extIndex] );
95 else
96 xInt[ i ] = xExt[extIndex];
97 }
98}
99
100void MinimTransformFunction::InvStepTransformation(const double * x, const double * sExt, double * sInt) const {
101 // inverse function transformation for steps (external -> internal)
102 for (unsigned int i = 0; i < NDim(); ++i ) {
103 unsigned int extIndex = fIndex[i];
104 const MinimTransformVariable & var = fVariables[ extIndex ];
105 assert ( !var.IsFixed() );
106 if (var.IsLimited() ) {
107 // bound variables
108 double x2 = x[extIndex] + sExt[extIndex];
109 if (var.HasUpperBound() && x2 >= var.UpperBound() )
110 x2 = x[extIndex] - sExt[extIndex];
111 // transform x and x2
112 double xint = var.ExternalToInternal ( x[extIndex] );
113 double x2int = var.ExternalToInternal( x2 );
114 sInt[i] = std::abs( x2int - xint);
115 }
116 else
117 sInt[ i ] = sExt[extIndex];
118 }
119}
120
121void MinimTransformFunction::GradientTransformation(const double * x, const double *gExt, double * gInt) const {
122 //transform gradient vector (external -> internal) at internal point x
123 unsigned int nfree = fIndex.size();
124 for (unsigned int i = 0; i < nfree; ++i ) {
125 unsigned int extIndex = fIndex[i];
126 const MinimTransformVariable & var = fVariables[ extIndex ];
127 assert (!var.IsFixed() );
128 if (var.IsLimited() )
129 gInt[i] = gExt[ extIndex ] * var.DerivativeIntToExt( x[i] );
130 else
131 gInt[i] = gExt[ extIndex ];
132 }
133}
134
135
136void MinimTransformFunction::MatrixTransformation(const double * x, const double *covInt, double * covExt) const {
137 //transform covariance matrix (internal -> external) at internal point x
138 // use row storages for matrices m(i,j) = rep[ i * dim + j]
139 // ignore fixed points
140 unsigned int nfree = fIndex.size();
141 unsigned int ntot = NTot();
142 for (unsigned int i = 0; i < nfree; ++i ) {
143 unsigned int iext = fIndex[i];
144 const MinimTransformVariable & ivar = fVariables[ iext ];
145 assert (!ivar.IsFixed());
146 double ddi = ( ivar.IsLimited() ) ? ivar.DerivativeIntToExt( x[i] ) : 1.0;
147 // loop on j variables for not fixed i variables (forget that matrix is symmetric) - could be optimized
148 for (unsigned int j = 0; j < nfree; ++j ) {
149 unsigned int jext = fIndex[j];
150 const MinimTransformVariable & jvar = fVariables[ jext ];
151 double ddj = ( jvar.IsLimited() ) ? jvar.DerivativeIntToExt( x[j] ) : 1.0;
152 assert (!jvar.IsFixed() );
153 covExt[ iext * ntot + jext] = ddi * ddj * covInt[ i * nfree + j];
154 }
155 }
156}
157
158
159 } // end namespace Math
160
161} // end namespace ROOT
162
#define f(i)
Definition: RSha256.hxx:104
static const double x2[5]
Interface (abstract class) for multi-dimensional functions providing a gradient calculation.
Definition: IFunction.h:327
unsigned int NDim() const
Retrieve the dimension of the function.
void InvTransformation(const double *xext, double *xint) const
inverse transformation (external -> internal)
MinimTransformFunction(const IMultiGradFunction *f, const std::vector< ROOT::Math::EMinimVariableType > &types, const std::vector< double > &values, const std::map< unsigned int, std::pair< double, double > > &bounds)
Constructor from a IMultiGradFunction interface (which is managed by the class) vector specifying the...
void GradientTransformation(const double *x, const double *gExt, double *gInt) const
transform gradient vector (external -> internal) at internal point x
void MatrixTransformation(const double *x, const double *covInt, double *covExt) const
transform covariance matrix (internal -> external) at internal point x use row storages for matrices ...
const double * Transformation(const double *x) const
transform from internal to external result is cached also inside the class
void InvStepTransformation(const double *x, const double *sext, double *sint) const
inverse transformation for steps (external -> internal) at external point x
std::vector< MinimTransformVariable > fVariables
MinimTransformVariable class Contains meta information of the variables such as bounds,...
Sin Transformation class for dealing with double bounded variables.
Sqrt Transformation class for dealing with lower bounded variables.
Sqrt Transformation class for dealing with upper bounded variables.
Double_t x[n]
Definition: legend1.C:17
Namespace for new Math classes and functions.
VSD Structures.
Definition: StringConv.hxx:21