/home/coin/svn-release/OptimizationSuite-1.1.0/CoinUtils/src/CoinFloatEqual.hpp File Reference

Function objects for testing equality of real numbers. More...

#include <algorithm>
#include <cmath>
#include "CoinFinite.hpp"

Include dependency graph for CoinFloatEqual.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	CoinAbsFltEq
	Equality to an absolute tolerance. More...
class	CoinRelFltEq
	Equality to a scaled tolerance. More...

---

Detailed Description

Function objects for testing equality of real numbers.

Two objects are provided; one tests for equality to an absolute tolerance, one to a scaled tolerance. The tests will handle IEEE floating point, but note that infinity == infinity. Mathematicians are rolling in their graves, but this matches the behaviour for the common practice of using DBL_MAX (numeric_limits<double>::max(), or similar large finite number) as infinity.

Example usage:

    double d1 = 3.14159 ;
    double d2 = d1 ;
    double d3 = d1+.0001 ;

    CoinAbsFltEq eq1 ;
    CoinAbsFltEq eq2(.001) ;

    assert(  eq1(d1,d2) ) ;
    assert( !eq1(d1,d3) ) ;
    assert(  eq2(d1,d3) ) ;
  

CoinRelFltEq follows the same pattern.

Definition in file CoinFloatEqual.hpp.