I must be doing something embarassingly stupid, but I can't find the bug
in the small attached program below. With the line marked "comment this
out" commented out, I get the expected number of solutions, but if I
leave this linear constraint uncommented, I get none (I expect just as
many solutions, but with a cost computed for each one).
#include "gecode/int.hh"
#include "gecode/search.hh"
#include "gecode/minimodel.hh"
#include <iostream>
using namespace std;
using namespace Gecode;
struct bug: public Space
{
IntVarArray X;
IntVar N;
bug(): X(*this, 4*4, 0, 1), N(*this, 0, 1000000)
{
Matrix<IntVarArray> MX(X, 4, 4);
for (int i=0; i<4; ++i)
{
linear(*this, MX.col(i), IRT_EQ, 1);
linear(*this, MX.row(i), IRT_EQ, 1);
}
IntArgs C(4*4);
Matrix<IntArgs> MC(C, 4, 4);
MC(0,0)=7; MC(0,1)=1; MC(0,2)=3; MC(0,3)=4;
MC(1,0)=8; MC(1,1)=2; MC(1,2)=5; MC(1,3)=1;
MC(2,0)=4; MC(2,1)=3; MC(2,2)=7; MC(2,3)=2;
MC(3,0)=3; MC(3,1)=1; MC(3,2)=6; MC(3,3)=8;
// comment this out
linear(*this, C, X, IRT_EQ, N);
branch(*this, X, INT_VAR_SIZE_MIN, INT_VAL_MIN);
}
bug(bool share, bug& b): Space(share, b)
{
X.update(*this, share, b.X);
N.update(*this, share, b.N);
}
virtual Space* copy(bool share)
{ return new bug(share, *this); }
void print() const
{
cout << "N = " << N << endl;
Matrix<IntVarArray> MX(X, 4, 4);
cout << MX << endl;
}
};
int main()
{
bug* model = new bug;
DFS<bug> engine(model);
delete model;
while (bug* solution = engine.next())
{
solution->print();
delete solution;
}
return 0;
}
Any help would be appreciated.
Cheers,
--Denys
_______________________________________________
Gecode users mailing list
[email protected]
https://www.gecode.org/mailman/listinfo/gecode-users
