| 1 | //
|
|---|
| 2 | // ********************************************************************
|
|---|
| 3 | // * License and Disclaimer *
|
|---|
| 4 | // * *
|
|---|
| 5 | // * The Geant4 software is copyright of the Copyright Holders of *
|
|---|
| 6 | // * the Geant4 Collaboration. It is provided under the terms and *
|
|---|
| 7 | // * conditions of the Geant4 Software License, included in the file *
|
|---|
| 8 | // * LICENSE and available at http://cern.ch/geant4/license . These *
|
|---|
| 9 | // * include a list of copyright holders. *
|
|---|
| 10 | // * *
|
|---|
| 11 | // * Neither the authors of this software system, nor their employing *
|
|---|
| 12 | // * institutes,nor the agencies providing financial support for this *
|
|---|
| 13 | // * work make any representation or warranty, express or implied, *
|
|---|
| 14 | // * regarding this software system or assume any liability for its *
|
|---|
| 15 | // * use. Please see the license in the file LICENSE and URL above *
|
|---|
| 16 | // * for the full disclaimer and the limitation of liability. *
|
|---|
| 17 | // * *
|
|---|
| 18 | // * This code implementation is the result of the scientific and *
|
|---|
| 19 | // * technical work of the GEANT4 collaboration. *
|
|---|
| 20 | // * By using, copying, modifying or distributing the software (or *
|
|---|
| 21 | // * any work based on the software) you agree to acknowledge its *
|
|---|
| 22 | // * use in resulting scientific publications, and indicate your *
|
|---|
| 23 | // * acceptance of all terms of the Geant4 Software license. *
|
|---|
| 24 | // ********************************************************************
|
|---|
| 25 | //
|
|---|
| 26 | // $Id: Gamma.cc,v 1.6 2006/06/29 19:14:28 gunter Exp $
|
|---|
| 27 | // GEANT4 tag $Name: geant4-09-04-beta-01 $
|
|---|
| 28 | //
|
|---|
| 29 | //
|
|---|
| 30 | // ------------------------------------------------------------
|
|---|
| 31 | // GEANT 4 class implementation
|
|---|
| 32 | // ------------------------------------------------------------
|
|---|
| 33 |
|
|---|
| 34 | #include <cmath>
|
|---|
| 35 | #include <string.h>
|
|---|
| 36 | #include "Gamma.hh"
|
|---|
| 37 |
|
|---|
| 38 | MyGamma::MyGamma(){}
|
|---|
| 39 |
|
|---|
| 40 | MyGamma::~MyGamma(){}
|
|---|
| 41 |
|
|---|
| 42 | //____________________________________________________________________________
|
|---|
| 43 | double MyGamma::Gamma(double z)
|
|---|
| 44 | {
|
|---|
| 45 | // Computation of gamma(z) for all z>0.
|
|---|
| 46 | //
|
|---|
| 47 | // The algorithm is based on the article by C.Lanczos [1] as denoted in
|
|---|
| 48 | // Numerical Recipes 2nd ed. on p. 207 (W.H.Press et al.).
|
|---|
| 49 | //
|
|---|
| 50 | // [1] C.Lanczos, SIAM Journal of Numerical Analysis B1 (1964), 86.
|
|---|
| 51 | //
|
|---|
| 52 | //--- Nve 14-nov-1998 UU-SAP Utrecht
|
|---|
| 53 |
|
|---|
| 54 | if (z<=0) return 0;
|
|---|
| 55 |
|
|---|
| 56 | double v = LnGamma(z);
|
|---|
| 57 | return std::exp(v);
|
|---|
| 58 | }
|
|---|
| 59 |
|
|---|
| 60 | //____________________________________________________________________________
|
|---|
| 61 | double MyGamma::Gamma(double a,double x)
|
|---|
| 62 | {
|
|---|
| 63 | // Computation of the incomplete gamma function P(a,x)
|
|---|
| 64 | //
|
|---|
| 65 | // The algorithm is based on the formulas and code as denoted in
|
|---|
| 66 | // Numerical Recipes 2nd ed. on p. 210-212 (W.H.Press et al.).
|
|---|
| 67 | //
|
|---|
| 68 | //--- Nve 14-nov-1998 UU-SAP Utrecht
|
|---|
| 69 |
|
|---|
| 70 | if (a <= 0 || x <= 0) return 0;
|
|---|
| 71 |
|
|---|
| 72 | if (x < (a+1)) return GamSer(a,x);
|
|---|
| 73 | else return GamCf(a,x);
|
|---|
| 74 | }
|
|---|
| 75 |
|
|---|
| 76 | //____________________________________________________________________________
|
|---|
| 77 | double MyGamma::GamCf(double a,double x)
|
|---|
| 78 | {
|
|---|
| 79 | // Computation of the incomplete gamma function P(a,x)
|
|---|
| 80 | // via its continued fraction representation.
|
|---|
| 81 | //
|
|---|
| 82 | // The algorithm is based on the formulas and code as denoted in
|
|---|
| 83 | // Numerical Recipes 2nd ed. on p. 210-212 (W.H.Press et al.).
|
|---|
| 84 | //
|
|---|
| 85 | //--- Nve 14-nov-1998 UU-SAP Utrecht
|
|---|
| 86 |
|
|---|
| 87 | int itmax = 100; // Maximum number of iterations
|
|---|
| 88 | double eps = 3.e-7; // Relative accuracy
|
|---|
| 89 | double fpmin = 1.e-30; // Smallest double value allowed here
|
|---|
| 90 |
|
|---|
| 91 | if (a <= 0 || x <= 0) return 0;
|
|---|
| 92 |
|
|---|
| 93 | double gln = LnGamma(a);
|
|---|
| 94 | double b = x+1-a;
|
|---|
| 95 | double c = 1/fpmin;
|
|---|
| 96 | double d = 1/b;
|
|---|
| 97 | double h = d;
|
|---|
| 98 | double an,del;
|
|---|
| 99 | for (int i=1; i<=itmax; i++) {
|
|---|
| 100 | an = double(-i)*(double(i)-a);
|
|---|
| 101 | b += 2;
|
|---|
| 102 | d = an*d+b;
|
|---|
| 103 | if (Abs(d) < fpmin) d = fpmin;
|
|---|
| 104 | c = b+an/c;
|
|---|
| 105 | if (Abs(c) < fpmin) c = fpmin;
|
|---|
| 106 | d = 1/d;
|
|---|
| 107 | del = d*c;
|
|---|
| 108 | h = h*del;
|
|---|
| 109 | if (Abs(del-1) < eps) break;
|
|---|
| 110 | //if (i==itmax) cout << "*GamCf(a,x)* a too large or itmax too small" << endl;
|
|---|
| 111 | }
|
|---|
| 112 | double v = Exp(-x+a*Log(x)-gln)*h;
|
|---|
| 113 | return (1-v);
|
|---|
| 114 | }
|
|---|
| 115 |
|
|---|
| 116 | //____________________________________________________________________________
|
|---|
| 117 | double MyGamma::GamSer(double a,double x)
|
|---|
| 118 | {
|
|---|
| 119 | // Computation of the incomplete gamma function P(a,x)
|
|---|
| 120 | // via its series representation.
|
|---|
| 121 | //
|
|---|
| 122 | // The algorithm is based on the formulas and code as denoted in
|
|---|
| 123 | // Numerical Recipes 2nd ed. on p. 210-212 (W.H.Press et al.).
|
|---|
| 124 | //
|
|---|
| 125 | //--- Nve 14-nov-1998 UU-SAP Utrecht
|
|---|
| 126 |
|
|---|
| 127 | int itmax = 100; // Maximum number of iterations
|
|---|
| 128 | double eps = 3.e-7; // Relative accuracy
|
|---|
| 129 |
|
|---|
| 130 | if (a <= 0 || x <= 0) return 0;
|
|---|
| 131 |
|
|---|
| 132 | double gln = LnGamma(a);
|
|---|
| 133 | double ap = a;
|
|---|
| 134 | double sum = 1/a;
|
|---|
| 135 | double del = sum;
|
|---|
| 136 | for (int n=1; n<=itmax; n++) {
|
|---|
| 137 | ap += 1;
|
|---|
| 138 | del = del*x/ap;
|
|---|
| 139 | sum += del;
|
|---|
| 140 | if (MyGamma::Abs(del) < Abs(sum*eps)) break;
|
|---|
| 141 | //if (n==itmax) cout << "*GamSer(a,x)* a too large or itmax too small" << endl;
|
|---|
| 142 | }
|
|---|
| 143 | double v = sum*Exp(-x+a*Log(x)-gln);
|
|---|
| 144 | return v;
|
|---|
| 145 | }
|
|---|
| 146 |
|
|---|
| 147 |
|
|---|
| 148 | double MyGamma::LnGamma(double z)
|
|---|
| 149 | {
|
|---|
| 150 | // Computation of ln[gamma(z)] for all z>0.
|
|---|
| 151 | //
|
|---|
| 152 | // The algorithm is based on the article by C.Lanczos [1] as denoted in
|
|---|
| 153 | // Numerical Recipes 2nd ed. on p. 207 (W.H.Press et al.).
|
|---|
| 154 | //
|
|---|
| 155 | // [1] C.Lanczos, SIAM Journal of Numerical Analysis B1 (1964), 86.
|
|---|
| 156 | //
|
|---|
| 157 | // The accuracy of the result is better than 2e-10.
|
|---|
| 158 | //
|
|---|
| 159 | //--- Nve 14-nov-1998 UU-SAP Utrecht
|
|---|
| 160 |
|
|---|
| 161 | if (z<=0) return 0;
|
|---|
| 162 |
|
|---|
| 163 | // Coefficients for the series expansion
|
|---|
| 164 | double c[7] = { 2.5066282746310005, 76.18009172947146, -86.50532032941677
|
|---|
| 165 | ,24.01409824083091, -1.231739572450155, 0.1208650973866179e-2
|
|---|
| 166 | ,-0.5395239384953e-5};
|
|---|
| 167 |
|
|---|
| 168 | double x = z;
|
|---|
| 169 | double y = x;
|
|---|
| 170 | double tmp = x+5.5;
|
|---|
| 171 | tmp = (x+0.5)*Log(tmp)-tmp;
|
|---|
| 172 | double ser = 1.000000000190015;
|
|---|
| 173 | for (int i=1; i<7; i++) {
|
|---|
| 174 | y += 1;
|
|---|
| 175 | ser += c[i]/y;
|
|---|
| 176 | }
|
|---|
| 177 | double v = tmp+Log(c[0]*ser/x);
|
|---|
| 178 | return v;
|
|---|
| 179 | }
|
|---|