On 3 Nov 2011, at 22:23, serge lemouton wrote: > Hello, > > I have slightly modified the tsp example to test the path constraint by > adding a "model" option to the script. > No solution is found with the -model path option : why ?
Have a look at the documentation for path: succ[e]==succ.size(), so the domains must be 0..succ.size() instead of 0..succ.size()-1. Cheers, Guido > Here is the modified code: > > <tsp2.cpp> > > Thanks for any hint, > > Yours, > > Serge > > On 25 oct. 2011, at 23:58, serge lemouton wrote: > >> Hello, >> >> I know that, and that is exactly what I was looking for ! >> But if it exists an hamiltonian circuit with n edges, it should exist at >> least n path with n-1 edges, right ? >> and I don't know how to search those with the Gecode "path" constraint ... >> Could you provide an example, please ? >> >> Thanks in advance, >> >> Serge >> >> >> >> On 25 oct. 2011, at 14:07, Christian Schulte wrote: >> >>> Hi, >>> >>> Please check the meaning of path again in the documentation: it enforces a >>> path and not a circuit! >>> >>> Christian >>> >>> -- >>> Christian Schulte, www.ict.kth.se/~cschulte/ >>> >>> >>> -----Original Message----- >>> From: [email protected] [mailto:[email protected]] On Behalf >>> Of serge lemouton >>> Sent: Saturday, October 22, 2011 5:59 PM >>> To: [email protected] >>> Subject: [gecode-users] looking for an example of path constraints >>> >>> Hi all, >>> >>> I am trying to use the new path constraints with the cost matrix, but >>> without any success, yet. >>> >>> In the TSP example script, why simply replacing the >>> >>> circuit(*this, c, succ, costs, total, opt.icl()); >>> by >>> path(*this, c, succ, start, end,costs, total, opt.icl()); >>> >>> does'nt find any solution ? >>> >>> Thanks for any hint, >>> >>> Serge Lemouton >>> _______________________________________________ >>> Gecode users mailing list >>> [email protected] >>> https://www.gecode.org/mailman/listinfo/gecode-users >>> >>> >> >> >> _______________________________________________ >> Gecode users mailing list >> [email protected] >> https://www.gecode.org/mailman/listinfo/gecode-users >> > > _______________________________________________ > Gecode users mailing list > [email protected] > https://www.gecode.org/mailman/listinfo/gecode-users -- Guido Tack, http://people.cs.kuleuven.be/~guido.tack/ _______________________________________________ Gecode users mailing list [email protected] https://www.gecode.org/mailman/listinfo/gecode-users
