Object-oriented Scientific Computing Library: Version 0.910
Protected Attributes
cern_adapt< func_t, nsub > Class Template Reference

Adaptive integration (CERNLIB) More...

#include <cern_adapt.h>

Inheritance diagram for cern_adapt< func_t, nsub >:
inte< func_t >

Detailed Description

template<class func_t = funct, size_t nsub = 100>
class cern_adapt< func_t, nsub >

Uses a base integration object (default is cern_gauss56) to perform adaptive integration by automatically subdividing the integration interval. At each step, the interval with the largest absolute uncertainty is divided in half. The routine succeeds if the absolute tolerance is less than tol_abs or if the relative tolerance is less than tol_rel, i.e.

\[ \mathrm{err}\leq\mathrm{tol\_abs}~\mathrm{or}~ \mathrm{err}\leq\mathrm{tol\_rel}\cdot|I| \]

where $ I $ is the current estimate for the integral and $ \mathrm{err} $ is the current estimate for the uncertainty. If the number of subdivisions exceeds the template parameter nsub, the error handler is called, since the integration may not have been successful. The number of subdivisions used in the last integration can be obtained from get_nsubdivisions().

The template parameter nsub, is the maximum number of subdivisions. It is automatically set to 100 in the original CERNLIB routine, and defaults to 100 here. The default base integration object is of type cern_gauss56. This is the CERNLIB default, but can be modified by calling set_inte().

This class is based on the CERNLIB routines RADAPT and DADAPT which are documented at http://wwwasdoc.web.cern.ch/wwwasdoc/shortwrupsdir/d102/top.html

Idea for Future:
  • Allow user to set the initial subdivisions?
  • It might be interesting to directly compare the performance of this class to gsl_inte_qag .
  • There is a fixme entry in the code which could be resolved.
  • Output the point where most subdividing was required?

Definition at line 71 of file cern_adapt.h.

Public Member Functions

Basic usage
virtual int integ_err (func_t &func, double a, double b, double &res, double &err)
 Integrate function func from a to b giving result res and error err.

Protected Attributes

double xlo [nsub]
 Lower end of subdivision.
double xhi [nsub]
 High end of subdivision.
double tval [nsub]
 Value of integral for subdivision.
double ters [nsub]
 Squared error for subdivision.
int prev_subdiv
 Previous number of subdivisions.
inte< func_t > * it
 The base integration object.

Integration object

cern_gauss56< func_t > def_inte
 Default integration object.
int set_inte (inte< func_t > &i)
 Set the base integration object to use.

Subdivisions

size_t nsubdiv
 Number of subdivisions.
size_t get_nsubdivisions ()
 Return the number of subdivisions used in the last integration.
int get_ith_subdivision (size_t i, double &xlow, double &xhigh, double &value, double &errsq)
 Return the ith subdivision.
template<class vec_t >
int get_subdivisions (vec_t &xlow, vec_t &xhigh, vec_t &value, vec_t &errsq)
 Return all of the subdivisions.

Field Documentation

template<class func_t = funct, size_t nsub = 100>
size_t cern_adapt< func_t, nsub >::nsubdiv

The options are

  • 0: Use previous binning and do not subdivide further
  • 1: Automatic - adapt until tolerance is attained (default)
  • n: (n>1) split first in n equal subdivisions, then adapt until tolerance is obtained.

Definition at line 266 of file cern_adapt.h.


The documentation for this class was generated from the following file:
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines

Documentation generated with Doxygen. Provided under the GNU Free Documentation License (see License Information).

Get Object-oriented Scientific Computing
Lib at SourceForge.net. Fast, secure and Free Open Source software
downloads.