[gecode-users] Restart search engine, adding a new constraint

Christian Schulte cschulte at kth.se
Thu Jun 6 12:02:25 CEST 2013


Hi,

It is spelled Cutoff. Please read the section in MPG about the master() and
slave() functions.

Best
Christian

--
Christian Schulte, www.ict.kth.se/~cschulte/


> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On
> Behalf Of Amina Kemmar
> Sent: Thursday, June 06, 2013 10:01 AM
> To: users at gecode.org
> Subject: [gecode-users] Restart search engine, adding a new constraint
> 
> On 06/01/2013 06:40 PM, Guido Tack wrote:
> > I guess the problem is that you only want to see x,y but you need to
search on
> z,k first? That's a bit more tricky.
> > You can use a restart search engine, which restarts after each solution.
Then
> for each solution you add the negation of the solution as a constraint in
the
> root, e.g. (not (x==1 and y==1)). I'm not sure how well this scales if you
have
> many solutions, but it could be worth a try.
> >
> > Cheers,
> > Guido
> >
> > On 31/05/2013, at 12:04, Christian Schulte <cschulte at kth.se> wrote:
> >
> >> Hi,
> >>
> >> I am not sure whether I understand your question correctly. But you
> >> can exchange the order of the search engine (that is, in my previous
> >> reply swap
> >> e1 and e2).
> >>
> >> Best
> >> Christian
> >>
> >> --
> >> Christian Schulte, Professor of Computer Science, KTH,
> >> www.ict.kth.se/~cschulte/
> >>
> >>
> >> -----Original Message-----
> >> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On
> >> Behalf Of Amina Kemmar
> >> Sent: Friday, May 31, 2013 10:00 AM
> >> To: users at gecode.org
> >> Subject: [gecode-users] Display the values of a set of variable
> >> without duplicates
> >>
> >> Hello,
> >>
> >> I already asked for the following problem:
> >>
> >> According to the variables defined in our model (x,y,z,k), I obtained
> >> the following solutions :
> >>   1,1,2,3
> >>   1,1,5,6
> >>   1,1,7,9
> >>   1.1,8,8
> >> I am interested to the values of the two variable x and y, so I want
> >> to avoid the enumeration of the three last solution in order to
> >> display just the first solution.
> >>
> >> I got the following response:
> >>
> >> One way to do it is as follows:
> >>   - You create a space s that only has a brancher for x and y
> >>   - You create a search engine e1 for s and find a solution, say s1
> >>   - You post a brancher in s1 for z and k.
> >>   - You create a search engine e2 for s and find just a single
solution!
> >>   - You can then also get more solutions from e1 and repeat the
> >> previous two steps.
> >>
> >> Thank you for the solution, it works very well, but I have an other
> >> problem which is the enumeration order of the variables x, y, z and
> >> k. I must begin the branching from the variables  z and k, then x, y
> >> (In this way, Gecode enumerates quickly the solutions but for the
> >> first order "x,y,z,k", it takes about 30 minutes to enumerate the first
> solution).
> >> Is there an other solution?
> >>
> >> Thank you.
> >> -----
> >> Amina
> >>
> >>
> >> _______________________________________________
> >> Gecode users mailing list
> >> users at gecode.org
> >> https://www.gecode.org/mailman/listinfo/gecode-users
> >>
> >>
> >> _______________________________________________
> >> Gecode users mailing list
> >> users at gecode.org
> >> https://www.gecode.org/mailman/listinfo/gecode-users
> Hi,
> 
> Thank you Guido for the response, this is exactly what I want to do, but I
red the
> documentation (section 9.4) , and I added the following code to restart
the
> search:
> 
> MyExample* s1 = new MyExample(opt);
> Search::Options o;
> Search::Cuttof* c = Search::Cuttof::constant(2); o.cuttof = c;
> RBS<DFS,MyExample> e3(s1, o);
> 
> I get some errors : 'Cuttof' is not a member of 'Gecode::Search'
> 
> My second question is how to add a new constraint, after the restart
search?
> 
> Thank you for your help.
> cheers.
> 
> I get
> 
> _______________________________________________
> Gecode users mailing list
> users at gecode.org
> https://www.gecode.org/mailman/listinfo/gecode-users




More information about the users mailing list