nojhan/eodev
Archived
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Adding sharing - needed to modify quite a few files, like eoDistance.h

Browse source 
make_algo_scalar.h and all related files, and the like
This commit is contained in:
evomarc 2004-05-25 08:03:30 +00:00
commit 795f63b7fe
14 changed files with 244 additions and 93 deletions
Download patch file Download diff file Expand all files Collapse all files

43
eo/src/utils/eoDistance.h
View file

@ -27,7 +27,6 @@
#define _eoDistance_H
#include <eoFunctor.h>
/**
This is a generic class for distance functors:
takes 2 things ane returns a double
@ -36,27 +35,63 @@ template< class EOT >
class eoDistance : public eoBF<const EOT &, const EOT &, double>
{};
/**
This is a generic class for Euclidain distance computation:
This is a generic class for Euclidain distance (L2 norm) computation:
assumes the 2 things are std::vectors of something that is double-castable
*/
template< class EOT >
class eoQuadDistance : public eoDistance<EOT>
{
public:
double operator()(const EOT & _v1, const EOT & _v2)
{
double sum=0.0;
double sum=0.0;
for (unsigned i=0; i<_v1.size(); i++)
{
double r = static_cast<double> (_v1[i]) - static_cast<double> (_v2[i]);
sum += r*r;
}
return sqrt(sum);
}
};
/**
This is a generic class for L1 distance computation:
assumes the 2 things are std::vectors of something
that is double-castable
For bitstrings, this is the Hamming distance
*/
template< class EOT >
class eoHammingDistance : public eoDistance<EOT>
{
public:
double operator()(const EOT & _v1, const EOT & _v2)
{
double sum=0.0;
for (unsigned i=0; i<_v1.size(); i++)
{
double r = static_cast<double> (_v1[i]) - static_cast<double> (_v2[i]);
sum += fabs(r);
}
return sum;
}
};
/* this distance measures the difference in fitness
* I am not sure it can be of any use, though ...
* except for some testing
*/
template< class EOT >
class eoFitnessDistance : public eoDistance<EOT>
{
public:
double operator()(const EOT & _v1, const EOT & _v2)
{
double d = _v1.fitness() - _v2.fitness();
return sqrt(d*d);
}
};