choco-users Mailing List for choco

Hi,

Yes, it seems there are errors if you use sparse domains and constraints=20
defined in extension. We will correct this problem...
Thank you for your help :)


-Guillaume


Varun Jain a =E9crit :
>
> Hi again,
>
> I thought I should give you my test code so that you can run it=20
> yourself. I have included the code at the bottom. Note that the CSP=20
> problem I was experimenting with has changed a bit (however, the=20
> strange behaviour still remains) Heres the new CSP problem:
>
> Variables:
> Variable0, Domain: {4, 2}
> Variable1, Domain: {5, 1, 4}
> Variable2, Domain: {3, 4, 6}
> Variable3, Domain: {1, 2}
> Variable4, Domain: {7, 8, 6, 1}
> Constraints:
> Constraint0, on variables 3 , Allowed tuples:
> 1
> Constraint1, on variables 2, 4 , Allowed tuples:
> 6, 7
> Constraint2, on variables 2 , Allowed tuples:
> 3
> 4
> Constraint3, on variables 4, 1 , Allowed tuples:
> 7, 1
> 8, 1
> 6, 5
> 1, 5
>
> Heres the output of running the program given below:
>
> Choco Solutions
> 2, 1, 3, 1, 8,
> 4, 1, 3, 1, 8,
>
> As you can see, none of the solutions satisfy constraint1. Still choco=20
> returns these solutions.
> Heres my code (BTW, Thanks for taking time out to help me.. I really=20
> appreciate it :-)):
>
> import choco.Problem;
> import choco.Constraint;
> import choco.integer.IntDomainVar;
> import choco.integer.IntVar;
> import java.util.ArrayList;
>
> /**
>  * @author Varun Jain (vj)
>  * @version $Revision$, $Date$
>  * @created Nov 17, 2006
>  */
> public class ChocoTest {
>   public static void main(String args[]) {
>     Problem chocoCSP =3D new Problem();
>     IntDomainVar[] vars =3D new IntDomainVar[5];
>     Constraint[] cons =3D new Constraint[4];
>     //-----Construct all variables
>     int id =3D 0;
>     vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new int[]{4, 2});
>     vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new int[]{5, 1, 4}=
);
>     vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new int[]{3, 4, 6}=
);
>     vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new int[]{1, 2});
>     vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new int[]{7, 8, 6,=
 1});
>     id =3D 0;
>     ArrayList<int[]> tuples;
>
>     //-----Now construct all constraints
>     //-----Constraint0
>     tuples =3D new ArrayList<int[]>();
>     tuples.add(new int[]{1});
>     cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[3]}, tuples, tr=
ue);
>     chocoCSP.post(cons[id++]);
>
>     //-----Constraint1
>     tuples =3D new ArrayList<int[]>();
>     tuples.add(new int[]{6, 7});
>     cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[2], vars[4]},=20
> tuples, true);
>     chocoCSP.post(cons[id++]);
>
>     //-----Constraint2
>     tuples =3D new ArrayList<int[]>();
>     tuples.add(new int[]{3});
>     tuples.add(new int[]{4});
>     cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[2]}, tuples, tr=
ue);
>     chocoCSP.post(cons[id++]);
>
>     //-----Constraint3
>     tuples =3D new ArrayList<int[]>();
>     tuples.add(new int[]{7, 1});
>     tuples.add(new int[]{8, 1});
>     tuples.add(new int[]{6, 5});
>     tuples.add(new int[]{1, 5});
>     cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[4], vars[1]},=20
> tuples, true);
>     chocoCSP.post(cons[id++]);
>
>     //-----Now get solutions
>     System.out.println("Choco Solutions");
>     if(chocoCSP.solve() =3D=3D Boolean.TRUE) {
>       do {
>         for(int i =3D 0; i < chocoCSP.getNbIntVars(); i ++) {
>          =20
> System.out.print(((IntDomainVar)chocoCSP.getIntVar(i)).getVal() + ", ")=
;
>         }
>         System.out.println();
>       } while(chocoCSP.nextSolution() =3D=3D Boolean.TRUE);
>     }
>   }
> }
>
>
>
>
>
>
> > Date: Sat, 18 Nov 2006 00:28:23 +0100
> > From: Gui...@em...
> > To: cho...@li...
> > Subject: Re: [Choco-users] solveAll() bug?
> >
> > Hi,
> >
> > This strange behavior is due to the fact that you use SolveAll() befo=
re
> > Solve(). SolveAll() and Solve()+nextSolution()'s loop are equivalents
> > (in other words you search for all the solutions twice), so usually w=
e
> > don't do that :) What do you want to do ?
> >
> > -Guillaume
> >
> > Varun Jain a =E9crit :
> > > Hi,
> > >
> > > Thank you very much for your quick reply. I guess the real question=
 I
> > > wanted to ask was that if choco could return results even when the
> > > problem is infeasible. I did my experiment with the following probl=
em:
> > >
> > > ------------------------------------------------
> > > Variables:
> > > Variable0, Domain: {3, 1, 4}
> > > Variable1, Domain: {2, 4, 6}
> > > Variable2, Domain: {2, 1}
> > > Variable3, Domain: {3, 1}
> > > Variable4, Domain: {6, 2, 5}
> > >
> > > Constraints:
> > > Constraint0, on variables 2 , Allowed values:
> > > 2
> > > Constraint1, on variables 2, 4 , Allowed values:
> > > 1, 6
> > > 2, 2
> > > Constraint2, on variables 4, 2 , Allowed values:
> > > 5, 2
> > > Constraint3, on variables 1, 3 , Allowed values:
> > > 2, 3
> > > ------------------------------------------------------
> > >
> > > All the constraints were created using Problem.makeTupleFC(IntVar[]
> > > varList, ArrayList tuples, boolean feasible). For example, the firs=
t
> > > constraint can be created as follows:
> > >
> > > ------------------------------------------
> > > Problem chocoCSP;
> > > /**
> > > * Set up the variables using chocoCSP.makeEnumIntVar()
> > > */
> > > IntVar[] varList =3D new IntVar[1];
> > > varList[0] =3D variables[2];
> > > ArrayList<int[]> tuples =3D new ArrayList<int[]>();
> > > tuples.add(new int[]{2});
> > > chocoCSP.makeTupleFC(varList, tuples, true);
> > > -------------------------------------------------
> > >
> > > To obtain solutions, I did the following:
> > >
> > > -----------------------------------------------
> > > System.out.println("SolveAll says: " + chocoCSP.solveAll());
> > > System.out.println("isFeasible says: " + chocoCSP.isFeasible());
> > > System.out.println("Choco Solutions");
> > > if(chocoCSP.solve() =3D=3D Boolean.TRUE) {
> > > do {
> > > for(int i =3D 0; i < chocoCSP.getNbIntVars(); i ++) {
> > > System.out.print(((IntDomainVar)
> > > chocoCSP.getIntVar(i)).getVal() + ", ");
> > > }
> > > System.out.println();
> > > } while(chocoCSP.nextSolution() =3D=3D Boolean.TRUE);
> > > }
> > > -----------------------------------------------
> > >
> > > Here is the output:
> > >
> > > SolveAll says: true
> > > isFeasible says: false
> > > Choco Solutions
> > > 1,2,2,3,5,
> > > 3,2,2,3,5,
> > > 4,2,2,3,5,
> > >
> > >
> > > Clearly, the solutions do not satisfy all the constraints. For
> > > instance, none of the solutions satisfy Constraint1. Is this expect=
ed
> > > behaviour??
> > >
> > > Thanks again,
> > > -Varun.
> > >
> > >
> > >
> > >
> > >=20
> -----------------------------------------------------------------------=
-
> > > > Date: Fri, 17 Nov 2006 11:31:40 +0100
> > > > From: Gui...@em...
> > > > To: jai...@ho...
> > > > Subject: Re: [Choco-users] solveAll() bug?
> > > >
> > > > Hi,
> > > >
> > > > I am not sure to understand what you try to do but i have some
> > > answers :)
> > > > -SolveAll still return true (it is hardcoded) but you can modify =
the
> > > > function solveAll() in choco.Problem
> > > >
> > > > In choco.Problem:
> > > > public Boolean solveAll() {
> > > > solver.firstSolution =3D false;
> > > > solver.generateSearchSolver(this);
> > > > solver.launch();
> > > > return Boolean.TRUE;
> > > > }
> > > >
> > > > -What kind of error do you have ? How do you implement the
> > > > nextSolution() loop ?
> > > > BTW, if you use nextSolution() to display solutions, you can use
> > > instead:
> > > >
> > > > "Solver.setVerbosity(Solver.SOLUTION);" before to begin to solve
> > > > and "Solver.flushLogs();" to print solutions
> > > >
> > > > else:
> > > >
> > > > if (pb.solve() =3D=3D Boolean.TRUE) {
> > > >
> > > > do {
> > > > //Code
> > > > }
> > > > } while(pb.nextSolution() =3D=3D Boolean.TRUE);
> > > > }
> > > >
> > > >
> > > > I hope it helps you... :)
> > > > -Guillaume
> > > >
> > > > Varun Jain a =E9crit :
> > > > > Hi..
> > > > > Is there any reason why the solveAll() method will return true
> > > even if
> > > > > there is no solution that satisfies all constraints. I was gett=
ing
> > > > > this behaviour so I did a little experiment: I set up a simple
> > > problem
> > > > > and executed the following:
> > > > >
> > > > > Problem chocoCSP;
> > > > > //...............
> > > > > //-----Set up the problem
> > > > > //...............
> > > > > System.out.println("SolveAll says: " + chocoCSP.solveAll());
> > > > > System.out.println("isFeasible says: " + chocoCSP.isFeasible())=
;
> > > > >
> > > > > The following was the output
> > > > >
> > > > > SolveAll says: true
> > > > > isFeasible says: false
> > > > >
> > > > > I thought solveAll() returns false if no solution exists.=20
> Also, for
> > > > > this problem, the solutions I get from the nextSolution()=20
> method do
> > > > > not satisfy all constraints. Any explainations would be really
> > > helpful?
> > > > >
> > > > > Thanks,
> > > > > -Varun.
> > > >
> > >=20
> -----------------------------------------------------------------------=
-
> > > Be one of the first to try Windows Live Mail.
> > >=20
> <http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-=
9b0e-4911fb2b2e6d>=20
>
> > >
> > >=20
> -----------------------------------------------------------------------=
-
> > >
> > >=20
> -----------------------------------------------------------------------=
--
> > > Take Surveys. Earn Cash. Influence the Future of IT
> > > Join SourceForge.net's Techsay panel and you'll get the chance to=20
> share your
> > > opinions on IT & business topics through brief surveys - and earn c=
ash
> > >=20
> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDEV
> > >=20
> -----------------------------------------------------------------------=
-
> > >
> > > _______________________________________________
> > > Choco-users mailing list
> > > Cho...@li...
> > > https://lists.sourceforge.net/lists/listinfo/choco-users
> > >
> >
> >
> >=20
> -----------------------------------------------------------------------=
--
> > Take Surveys. Earn Cash. Influence the Future of IT
> > Join SourceForge.net's Techsay panel and you'll get the chance to=20
> share your
> > opinions on IT & business topics through brief surveys - and earn cas=
h
> >=20
> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDEV
> > _______________________________________________
> > Choco-users mailing list
> > Cho...@li...
> > https://lists.sourceforge.net/lists/listinfo/choco-users
>
> -----------------------------------------------------------------------=
-
> Be one of the first to try Windows Live Mail.=20
> <http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-=
9b0e-4911fb2b2e6d>=20
>
> -----------------------------------------------------------------------=
-
>
> -----------------------------------------------------------------------=
--
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share=
 your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDEV
> -----------------------------------------------------------------------=
-
>
> _______________________________________________
> Choco-users mailing list
> Cho...@li...
> https://lists.sourceforge.net/lists/listinfo/choco-users
>  =20

Hi again,I thought I should give you my test code so that you can run it yo=
urself. I have included the code at the bottom. Note that the CSP problem I=
 was experimenting with has changed a bit (however, the strange behaviour s=
till remains) Heres the new CSP problem:Variables:Variable0, Domain: {4, 2}=
Variable1, Domain: {5, 1, 4}Variable2, Domain: {3, 4, 6}Variable3, Domain: =
{1, 2}Variable4, Domain: {7, 8, 6, 1}Constraints:Constraint0, on variables =
3 , Allowed tuples:1Constraint1, on variables 2, 4 , Allowed tuples:6, 7Con=
straint2, on variables 2 , Allowed tuples:34Constraint3, on variables 4, 1 =
, Allowed tuples:7, 18, 16, 51, 5Heres the output of running the program gi=
ven below:Choco Solutions2, 1, 3, 1, 8, 4, 1, 3, 1, 8, As you can see, none=
 of the solutions satisfy constraint1. Still choco returns these solutions.=
Heres my code (BTW, Thanks for taking time out to help me.. I really apprec=
iate it :-)):import choco.Problem;import choco.Constraint;import choco.inte=
ger.IntDomainVar;import choco.integer.IntVar;import java.util.ArrayList;/**=
 * @author Varun Jain (vj) * @version $Revision$, $Date$ * @created Nov 17,=
 2006 */public class ChocoTest {  public static void main(String args[]) { =
   Problem chocoCSP =3D new Problem();    IntDomainVar[] vars =3D new IntDo=
mainVar[5];    Constraint[] cons =3D new Constraint[4];    //-----Construct=
 all variables    int id =3D 0;    vars[id] =3D chocoCSP.makeEnumIntVar("v"=
 + id++, new int[]{4, 2});    vars[id] =3D chocoCSP.makeEnumIntVar("v" + id=
++, new int[]{5, 1, 4});    vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++=
, new int[]{3, 4, 6});    vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, =
new int[]{1, 2});    vars[id] =3D chocoCSP.makeEnumIntVar("v" + id++, new i=
nt[]{7, 8, 6, 1});    id =3D 0;    ArrayList<int[]> tuples;    //-----Now c=
onstruct all constraints    //-----Constraint0    tuples =3D new ArrayList<=
int[]>();    tuples.add(new int[]{1});    cons[id] =3D chocoCSP.makeTupleFC=
(new IntVar[]{vars[3]}, tuples, true);    chocoCSP.post(cons[id++]);    //-=
----Constraint1=0A=
    tuples =3D new ArrayList<int[]>();    tuples.add(new int[]{6, 7});    c=
ons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[2], vars[4]}, tuples, tr=
ue);    chocoCSP.post(cons[id++]);    //-----Constraint2=0A=
    tuples =3D new ArrayList<int[]>();    tuples.add(new int[]{3});    tupl=
es.add(new int[]{4});    cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{var=
s[2]}, tuples, true);    chocoCSP.post(cons[id++]);    //-----Constraint3=
=0A=
    tuples =3D new ArrayList<int[]>();    tuples.add(new int[]{7, 1});    t=
uples.add(new int[]{8, 1});    tuples.add(new int[]{6, 5});    tuples.add(n=
ew int[]{1, 5});    cons[id] =3D chocoCSP.makeTupleFC(new IntVar[]{vars[4],=
 vars[1]}, tuples, true);    chocoCSP.post(cons[id++]);    //-----Now get s=
olutions    System.out.println("Choco Solutions");    if(chocoCSP.solve() =
=3D=3D Boolean.TRUE) {      do {        for(int i =3D 0; i < chocoCSP.getNb=
IntVars(); i ++) {          System.out.print(((IntDomainVar)chocoCSP.getInt=
Var(i)).getVal() + ", ");        }        System.out.println();      } whil=
e(chocoCSP.nextSolution() =3D=3D Boolean.TRUE);    }  }}> Date: Sat, 18 Nov=
 2006 00:28:23 +0100> From: Gui...@em...> To: choco-users@lists=
.sourceforge.net> Subject: Re: [Choco-users] solveAll() bug?> > Hi,> > This=
 strange behavior is due to the fact that you use SolveAll() before > Solve=
().  SolveAll() and  Solve()+nextSolution()'s loop  are equivalents > (in o=
ther words you search for all the solutions  twice), so usually we > don't =
do that :) What do you want to do ?> > -Guillaume> > Varun Jain a =E9crit :=
> > Hi,> >> > Thank you very much for your quick reply. I guess the real qu=
estion I > > wanted to ask was that if choco could return results even when=
 the > > problem is infeasible. I did my experiment with the following prob=
lem:> >> > ------------------------------------------------> > Variables:> =
> Variable0, Domain: {3, 1, 4}> > Variable1, Domain: {2, 4, 6}> > Variable2=
, Domain: {2, 1}> > Variable3, Domain: {3, 1}> > Variable4, Domain: {6, 2, =
5}> >> > Constraints:> > Constraint0, on variables 2 , Allowed values:> > 2=
> > Constraint1, on variables 2, 4 , Allowed values:> > 1, 6> > 2, 2> > Con=
straint2, on variables 4, 2 , Allowed values:> > 5, 2> > Constraint3, on va=
riables 1, 3 , Allowed values:> > 2, 3> > ---------------------------------=
---------------------> >> > All the constraints were created using Problem.=
makeTupleFC(IntVar[] > > varList, ArrayList tuples, boolean feasible). For =
example, the first > > constraint can be created as follows:> >> > --------=
----------------------------------> > Problem chocoCSP;> > /**> >  * Set up=
 the variables using chocoCSP.makeEnumIntVar()> >  */> > IntVar[] varList =
=3D new IntVar[1];> > varList[0] =3D variables[2];> > ArrayList<int[]> tupl=
es =3D new ArrayList<int[]>();> > tuples.add(new int[]{2});> > chocoCSP.mak=
eTupleFC(varList, tuples, true);> > ---------------------------------------=
----------> >> > To obtain solutions, I did the following:> >> > ----------=
-------------------------------------> >     System.out.println("SolveAll s=
ays: " + chocoCSP.solveAll());> >     System.out.println("isFeasible says: =
" + chocoCSP.isFeasible());> >     System.out.println("Choco Solutions");> =
>     if(chocoCSP.solve() =3D=3D Boolean.TRUE) {> >       do {> >         f=
or(int i =3D 0; i < chocoCSP.getNbIntVars(); i ++) {> >           System.ou=
t.print(((IntDomainVar) > > chocoCSP.getIntVar(i)).getVal() + ", ");> >    =
     }> >         System.out.println();> >       } while(chocoCSP.nextSolut=
ion() =3D=3D Boolean.TRUE);> >     }> > -----------------------------------=
------------> >> > Here is the output:> >> > SolveAll says: true> > isFeasi=
ble says: false> > Choco Solutions> > 1,2,2,3,5,> > 3,2,2,3,5,> > 4,2,2,3,5=
,> >> >> > Clearly, the solutions do not satisfy all the constraints. For >=
 > instance, none of the solutions satisfy Constraint1. Is this expected > =
> behaviour??> >> > Thanks again,> > -Varun.> >> >> >> >> > ---------------=
---------------------------------------------------------> > > Date: Fri, 1=
7 Nov 2006 11:31:40 +0100> > > From: Gui...@em...> > > To: jain=
_v...@ho...> > > Subject: Re: [Choco-users] solveAll() bug?> > >> > =
> Hi,> > >> > > I am not sure to understand what you try to do but i have s=
ome > > answers :)> > > -SolveAll still return true (it is hardcoded) but y=
ou can modify the> > > function solveAll() in choco.Problem> > >> > > In ch=
oco.Problem:> > > public Boolean solveAll() {> > > solver.firstSolution =3D=
 false;> > > solver.generateSearchSolver(this);> > > solver.launch();> > > =
return Boolean.TRUE;> > > }> > >> > > -What kind of error do you have ? How=
 do you implement the> > > nextSolution() loop ?> > > BTW, if you use nextS=
olution() to display solutions, you can use > > instead:> > >> > > "Solver.=
setVerbosity(Solver.SOLUTION);" before to begin to solve> > > and "Solver.f=
lushLogs();" to print solutions> > >> > > else:> > >> > > if (pb.solve() =
=3D=3D Boolean.TRUE) {> > >> > > do {> > > //Code> > > }> > > } while(pb.ne=
xtSolution() =3D=3D Boolean.TRUE);> > > }> > >> > >> > > I hope it helps yo=
u... :)> > > -Guillaume> > >> > > Varun Jain a =E9crit :> > > > Hi..> > > >=
 Is there any reason why the solveAll() method will return true > > even if=
> > > > there is no solution that satisfies all constraints. I was getting>=
 > > > this behaviour so I did a little experiment: I set up a simple > > p=
roblem> > > > and executed the following:> > > >> > > > Problem chocoCSP;> =
> > > //...............> > > > //-----Set up the problem> > > > //.........=
......> > > > System.out.println("SolveAll says: " + chocoCSP.solveAll());>=
 > > > System.out.println("isFeasible says: " + chocoCSP.isFeasible());> > =
> >> > > > The following was the output> > > >> > > > SolveAll says: true> =
> > > isFeasible says: false> > > >> > > > I thought solveAll() returns fal=
se if no solution exists. Also, for> > > > this problem, the solutions I ge=
t from the nextSolution() method do> > > > not satisfy all constraints. Any=
 explainations would be really > > helpful?> > > >> > > > Thanks,> > > > -V=
arun.> > >> > -------------------------------------------------------------=
-----------> > Be one of the first to try Windows Live Mail. > > <http://id=
eas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-4911fb2b2=
e6d> > >> > ---------------------------------------------------------------=
---------> >> > -----------------------------------------------------------=
--------------> > Take Surveys. Earn Cash. Influence the Future of IT> > Jo=
in SourceForge.net's Techsay panel and you'll get the chance to share your>=
 > opinions on IT & business topics through brief surveys - and earn cash> =
> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDEV> > -----------------------------------------------------------------=
-------> >> > _______________________________________________> > Choco-user=
s mailing list> > Cho...@li...> > https://lists.source=
forge.net/lists/listinfo/choco-users> >   > > > ---------------------------=
----------------------------------------------> Take Surveys. Earn Cash. In=
fluence the Future of IT> Join SourceForge.net's Techsay panel and you'll g=
et the chance to share your> opinions on IT & business topics through brief=
 surveys - and earn cash> http://www.techsay.com/default.php?page=3Djoin.ph=
p&p=3Dsourceforge&CID=3DDEVDEV> ___________________________________________=
____> Choco-users mailing list> Cho...@li...> https://=
lists.sourceforge.net/lists/listinfo/choco-users
_________________________________________________________________
Be one of the first to try Windows Live Mail.
http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-=
4911fb2b2e6d=

Hi,Thanks again for your reply. Actually I included SolveAll() later in an =
attempt to try and understand whats wrong. Basically, I wanted to extract a=
ll solutions if any exists. So I did the following:    System.out.println("=
Choco Solutions");    if(chocoCSP.solve() =3D=3D Boolean.TRUE) {      do { =
       for(int i =3D 0; i < chocoCSP.getNbIntVars(); i ++) {          Syste=
m.out.print(((IntDomainVar)chocoCSP.getIntVar(i)).getVal() + ", ");        =
}        System.out.println();      } while(chocoCSP.nextSolution() =3D=3D =
Boolean.TRUE);    }And the output was still the following:Choco Solutions1,=
2,2,3,5,3,2,2,3,5,4,2,2,3,5,I included the SolveAll() and isFeasible() meth=
ods later since the above was not working. So even in the absence of SolveA=
ll() I get the strange behaviour.-Varun.> Date: Sat, 18 Nov 2006 00:28:23 +=
0100> From: Gui...@em...> To: cho...@li...=
> Subject: Re: [Choco-users] solveAll() bug?> > Hi,> > This strange behavio=
r is due to the fact that you use SolveAll() before > Solve().  SolveAll() =
and  Solve()+nextSolution()'s loop  are equivalents > (in other words you s=
earch for all the solutions  twice), so usually we > don't do that :) What =
do you want to do ?> > -Guillaume> > Varun Jain a =E9crit :> > Hi,> >> > Th=
ank you very much for your quick reply. I guess the real question I > > wan=
ted to ask was that if choco could return results even when the > > problem=
 is infeasible. I did my experiment with the following problem:> >> > -----=
-------------------------------------------> > Variables:> > Variable0, Dom=
ain: {3, 1, 4}> > Variable1, Domain: {2, 4, 6}> > Variable2, Domain: {2, 1}=
> > Variable3, Domain: {3, 1}> > Variable4, Domain: {6, 2, 5}> >> > Constra=
ints:> > Constraint0, on variables 2 , Allowed values:> > 2> > Constraint1,=
 on variables 2, 4 , Allowed values:> > 1, 6> > 2, 2> > Constraint2, on var=
iables 4, 2 , Allowed values:> > 5, 2> > Constraint3, on variables 1, 3 , A=
llowed values:> > 2, 3> > -------------------------------------------------=
-----> >> > All the constraints were created using Problem.makeTupleFC(IntV=
ar[] > > varList, ArrayList tuples, boolean feasible). For example, the fir=
st > > constraint can be created as follows:> >> > ------------------------=
------------------> > Problem chocoCSP;> > /**> >  * Set up the variables u=
sing chocoCSP.makeEnumIntVar()> >  */> > IntVar[] varList =3D new IntVar[1]=
;> > varList[0] =3D variables[2];> > ArrayList<int[]> tuples =3D new ArrayL=
ist<int[]>();> > tuples.add(new int[]{2});> > chocoCSP.makeTupleFC(varList,=
 tuples, true);> > -------------------------------------------------> >> > =
To obtain solutions, I did the following:> >> > ---------------------------=
--------------------> >     System.out.println("SolveAll says: " + chocoCSP=
.solveAll());> >     System.out.println("isFeasible says: " + chocoCSP.isFe=
asible());> >     System.out.println("Choco Solutions");> >     if(chocoCSP=
.solve() =3D=3D Boolean.TRUE) {> >       do {> >         for(int i =3D 0; i=
 < chocoCSP.getNbIntVars(); i ++) {> >           System.out.print(((IntDoma=
inVar) > > chocoCSP.getIntVar(i)).getVal() + ", ");> >         }> >        =
 System.out.println();> >       } while(chocoCSP.nextSolution() =3D=3D Bool=
ean.TRUE);> >     }> > -----------------------------------------------> >> =
> Here is the output:> >> > SolveAll says: true> > isFeasible says: false> =
> Choco Solutions> > 1,2,2,3,5,> > 3,2,2,3,5,> > 4,2,2,3,5,> >> >> > Clearl=
y, the solutions do not satisfy all the constraints. For > > instance, none=
 of the solutions satisfy Constraint1. Is this expected > > behaviour??> >>=
 > Thanks again,> > -Varun.> >> >> >> >> > --------------------------------=
----------------------------------------> > > Date: Fri, 17 Nov 2006 11:31:=
40 +0100> > > From: Gui...@em...> > > To: jai...@ho...=
m> > > Subject: Re: [Choco-users] solveAll() bug?> > >> > > Hi,> > >> > > I=
 am not sure to understand what you try to do but i have some > > answers :=
)> > > -SolveAll still return true (it is hardcoded) but you can modify the=
> > > function solveAll() in choco.Problem> > >> > > In choco.Problem:> > >=
 public Boolean solveAll() {> > > solver.firstSolution =3D false;> > > solv=
er.generateSearchSolver(this);> > > solver.launch();> > > return Boolean.TR=
UE;> > > }> > >> > > -What kind of error do you have ? How do you implement=
 the> > > nextSolution() loop ?> > > BTW, if you use nextSolution() to disp=
lay solutions, you can use > > instead:> > >> > > "Solver.setVerbosity(Solv=
er.SOLUTION);" before to begin to solve> > > and "Solver.flushLogs();" to p=
rint solutions> > >> > > else:> > >> > > if (pb.solve() =3D=3D Boolean.TRUE=
) {> > >> > > do {> > > //Code> > > }> > > } while(pb.nextSolution() =3D=3D=
 Boolean.TRUE);> > > }> > >> > >> > > I hope it helps you... :)> > > -Guill=
aume> > >> > > Varun Jain a =E9crit :> > > > Hi..> > > > Is there any reaso=
n why the solveAll() method will return true > > even if> > > > there is no=
 solution that satisfies all constraints. I was getting> > > > this behavio=
ur so I did a little experiment: I set up a simple > > problem> > > > and e=
xecuted the following:> > > >> > > > Problem chocoCSP;> > > > //...........=
....> > > > //-----Set up the problem> > > > //...............> > > > Syste=
m.out.println("SolveAll says: " + chocoCSP.solveAll());> > > > System.out.p=
rintln("isFeasible says: " + chocoCSP.isFeasible());> > > >> > > > The foll=
owing was the output> > > >> > > > SolveAll says: true> > > > isFeasible sa=
ys: false> > > >> > > > I thought solveAll() returns false if no solution e=
xists. Also, for> > > > this problem, the solutions I get from the nextSolu=
tion() method do> > > > not satisfy all constraints. Any explainations woul=
d be really > > helpful?> > > >> > > > Thanks,> > > > -Varun.> > >> > -----=
-------------------------------------------------------------------> > Be o=
ne of the first to try Windows Live Mail. > > <http://ideas.live.com/progra=
mpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-4911fb2b2e6d> > >> > -------=
-----------------------------------------------------------------> >> > ---=
----------------------------------------------------------------------> > T=
ake Surveys. Earn Cash. Influence the Future of IT> > Join SourceForge.net'=
s Techsay panel and you'll get the chance to share your> > opinions on IT &=
 business topics through brief surveys - and earn cash> > http://www.techsa=
y.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3DDEVDEV> > ---------=
---------------------------------------------------------------> >> > _____=
__________________________________________> > Choco-users mailing list> > C=
hoc...@li...> > https://lists.sourceforge.net/lists/lis=
tinfo/choco-users> >   > > > ----------------------------------------------=
---------------------------> Take Surveys. Earn Cash. Influence the Future =
of IT> Join SourceForge.net's Techsay panel and you'll get the chance to sh=
are your> opinions on IT & business topics through brief surveys - and earn=
 cash> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&C=
ID=3DDEVDEV> _______________________________________________> Choco-users m=
ailing list> Cho...@li...> https://lists.sourceforge.n=
et/lists/listinfo/choco-users
_________________________________________________________________
Be one of the first to try Windows Live Mail.
http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-=
4911fb2b2e6d=

Hi,

This strange behavior is due to the fact that you use SolveAll() before=20
Solve().  SolveAll() and  Solve()+nextSolution()'s loop  are equivalents=20
(in other words you search for all the solutions  twice), so usually we=20
don't do that :) What do you want to do ?

-Guillaume

Varun Jain a =E9crit :
> Hi,
>
> Thank you very much for your quick reply. I guess the real question I=20
> wanted to ask was that if choco could return results even when the=20
> problem is infeasible. I did my experiment with the following problem:
>
> ------------------------------------------------
> Variables:
> Variable0, Domain: {3, 1, 4}
> Variable1, Domain: {2, 4, 6}
> Variable2, Domain: {2, 1}
> Variable3, Domain: {3, 1}
> Variable4, Domain: {6, 2, 5}
>
> Constraints:
> Constraint0, on variables 2 , Allowed values:
> 2
> Constraint1, on variables 2, 4 , Allowed values:
> 1, 6
> 2, 2
> Constraint2, on variables 4, 2 , Allowed values:
> 5, 2
> Constraint3, on variables 1, 3 , Allowed values:
> 2, 3
> ------------------------------------------------------
>
> All the constraints were created using Problem.makeTupleFC(IntVar[]=20
> varList, ArrayList tuples, boolean feasible). For example, the first=20
> constraint can be created as follows:
>
> ------------------------------------------
> Problem chocoCSP;
> /**
>  * Set up the variables using chocoCSP.makeEnumIntVar()
>  */
> IntVar[] varList =3D new IntVar[1];
> varList[0] =3D variables[2];
> ArrayList<int[]> tuples =3D new ArrayList<int[]>();
> tuples.add(new int[]{2});
> chocoCSP.makeTupleFC(varList, tuples, true);
> -------------------------------------------------
>
> To obtain solutions, I did the following:
>
> -----------------------------------------------
>     System.out.println("SolveAll says: " + chocoCSP.solveAll());
>     System.out.println("isFeasible says: " + chocoCSP.isFeasible());
>     System.out.println("Choco Solutions");
>     if(chocoCSP.solve() =3D=3D Boolean.TRUE) {
>       do {
>         for(int i =3D 0; i < chocoCSP.getNbIntVars(); i ++) {
>           System.out.print(((IntDomainVar)=20
> chocoCSP.getIntVar(i)).getVal() + ", ");
>         }
>         System.out.println();
>       } while(chocoCSP.nextSolution() =3D=3D Boolean.TRUE);
>     }
> -----------------------------------------------
>
> Here is the output:
>
> SolveAll says: true
> isFeasible says: false
> Choco Solutions
> 1,2,2,3,5,
> 3,2,2,3,5,
> 4,2,2,3,5,
>
>
> Clearly, the solutions do not satisfy all the constraints. For=20
> instance, none of the solutions satisfy Constraint1. Is this expected=20
> behaviour??
>
> Thanks again,
> -Varun.
>
>
>
>
> -----------------------------------------------------------------------=
-
> > Date: Fri, 17 Nov 2006 11:31:40 +0100
> > From: Gui...@em...
> > To: jai...@ho...
> > Subject: Re: [Choco-users] solveAll() bug?
> >
> > Hi,
> >
> > I am not sure to understand what you try to do but i have some=20
> answers :)
> > -SolveAll still return true (it is hardcoded) but you can modify the
> > function solveAll() in choco.Problem
> >
> > In choco.Problem:
> > public Boolean solveAll() {
> > solver.firstSolution =3D false;
> > solver.generateSearchSolver(this);
> > solver.launch();
> > return Boolean.TRUE;
> > }
> >
> > -What kind of error do you have ? How do you implement the
> > nextSolution() loop ?
> > BTW, if you use nextSolution() to display solutions, you can use=20
> instead:
> >
> > "Solver.setVerbosity(Solver.SOLUTION);" before to begin to solve
> > and "Solver.flushLogs();" to print solutions
> >
> > else:
> >
> > if (pb.solve() =3D=3D Boolean.TRUE) {
> >
> > do {
> > //Code
> > }
> > } while(pb.nextSolution() =3D=3D Boolean.TRUE);
> > }
> >
> >
> > I hope it helps you... :)
> > -Guillaume
> >
> > Varun Jain a =E9crit :
> > > Hi..
> > > Is there any reason why the solveAll() method will return true=20
> even if
> > > there is no solution that satisfies all constraints. I was getting
> > > this behaviour so I did a little experiment: I set up a simple=20
> problem
> > > and executed the following:
> > >
> > > Problem chocoCSP;
> > > //...............
> > > //-----Set up the problem
> > > //...............
> > > System.out.println("SolveAll says: " + chocoCSP.solveAll());
> > > System.out.println("isFeasible says: " + chocoCSP.isFeasible());
> > >
> > > The following was the output
> > >
> > > SolveAll says: true
> > > isFeasible says: false
> > >
> > > I thought solveAll() returns false if no solution exists. Also, for
> > > this problem, the solutions I get from the nextSolution() method do
> > > not satisfy all constraints. Any explainations would be really=20
> helpful?
> > >
> > > Thanks,
> > > -Varun.
> >
> -----------------------------------------------------------------------=
-
> Be one of the first to try Windows Live Mail.=20
> <http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-=
9b0e-4911fb2b2e6d>=20
>
> -----------------------------------------------------------------------=
-
>
> -----------------------------------------------------------------------=
--
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share=
 your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDEV
> -----------------------------------------------------------------------=
-
>
> _______________________________________________
> Choco-users mailing list
> Cho...@li...
> https://lists.sourceforge.net/lists/listinfo/choco-users
>  =20

=0A=
=0A=
=0A=
=0A=
=0A=
Hi,Thank you very much for your quick reply. I=0A=
guess the real question I wanted to ask was that if choco could return=0A=
results even when the problem is infeasible. I did my experiment with=0A=
the following problem:------------------------------------------------Varia=
bles:Variable0, Domain: {3, 1, 4}Variable1, Domain: {2, 4, 6}Variable2, Dom=
ain: {2, 1}Variable3, Domain: {3, 1}Variable4, Domain: {6, 2, 5}Constraints=
:Constraint0, on variables 2 , Allowed values:2Constraint1, on variables 2,=
 4 , Allowed values:1, 62, 2Constraint2, on variables 4, 2 , Allowed values=
:5, 2Constraint3, on variables 1, 3 , Allowed values:2, 3------------------=
------------------------------------All=0A=
the constraints were created using Problem.makeTupleFC(IntVar[]=0A=
varList, ArrayList tuples, boolean feasible). For example, the first=0A=
constraint can be created as follows:--------------------------------------=
----Problem chocoCSP;/** * Set up the variables using chocoCSP.makeEnumIntV=
ar() */IntVar[] varList =3D new IntVar[1];varList[0] =3D variables[2];Array=
List<int[]> tuples =3D new ArrayList<int[]>();tuples.add(new int[]{2});choc=
oCSP.makeTupleFC(varList, tuples, true);-----------------------------------=
--------------To obtain solutions, I did the following:--------------------=
---------------------------    System.out.println("SolveAll says: " + choco=
CSP.solveAll());    System.out.println("isFeasible says: " + chocoCSP.isFea=
sible());    System.out.println("Choco Solutions");    if(chocoCSP.solve() =
=3D=3D Boolean.TRUE) {      do {        for(int i =3D 0; i < chocoCSP.getNb=
IntVars(); i ++) {          System.out.print(((IntDomainVar) chocoCSP.getIn=
tVar(i)).getVal() + ", ");        }        System.out.println();      } whi=
le(chocoCSP.nextSolution() =3D=3D Boolean.TRUE);    }----------------------=
-------------------------Here is the output:SolveAll says: trueisFeasible s=
ays: falseChoco Solutions1,2,2,3,5,3,2,2,3,5,4,2,2,3,5,Clearly,=0A=
the solutions do not satisfy all the constraints. For instance, none of=0A=
the solutions satisfy Constraint1. Is this expected behaviour??Thanks again=
,-Varun.> Date: Fri, 17 Nov 2006 11:31:40 +0100> From: Guillaume.Richaud@em=
n.fr> To: jai...@ho...> Subject: Re: [Choco-users] solveAll() bug=
?> > Hi,> > I am not sure to understand what you try to do but i have some =
answers :)> -SolveAll still return true (it is hardcoded) but you can modif=
y the > function solveAll() in choco.Problem> > In choco.Problem:>   public=
 Boolean solveAll() {>     solver.firstSolution =3D false;>     solver.gene=
rateSearchSolver(this);>     solver.launch();>     return Boolean.TRUE;>   =
}> > -What kind of error do you have ? How do you implement the > nextSolut=
ion() loop ?> BTW, if you use nextSolution() to display solutions, you can =
use instead:> > "Solver.setVerbosity(Solver.SOLUTION);" before to begin to =
solve> and "Solver.flushLogs();" to print solutions> > else:> > if (pb.solv=
e() =3D=3D Boolean.TRUE) {> >   do {> 	//Code>     }>   } while(pb.nextSolu=
tion() =3D=3D Boolean.TRUE);> }> > > I hope it helps you... :)> -Guillaume>=
 > Varun Jain a =E9crit :> > Hi..> > Is there any reason why the solveAll()=
 method will return true even if > > there is no solution that satisfies al=
l constraints. I was getting > > this behaviour so I did a little experimen=
t: I set up a simple problem > > and executed the following: > >  > > Probl=
em chocoCSP; > > //............... > > //-----Set up the problem > > //....=
........... > > System.out.println("SolveAll says: " + chocoCSP.solveAll())=
; > > System.out.println("isFeasible says: " + chocoCSP.isFeasible()); > > =
 > > The following was the output > >  > > SolveAll says: true > > isFeasib=
le says: false > >  > > I thought solveAll() returns false if no solution e=
xists. Also, for > > this problem, the solutions I get from the nextSolutio=
n() method do > > not satisfy all constraints. Any explainations would be r=
eally helpful? > >  > > Thanks, > > -Varun. >=20
_________________________________________________________________
Be one of the first to try Windows Live Mail.
http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-=
4911fb2b2e6d=

Hi..Is there any reason why the solveAll() method will return true even if=
=0A=
there is no solution that satisfies all constraints. I was getting=0A=
this behaviour so I did a little experiment: I set up a simple problem=0A=
and executed the following: =0A=
 =0A=
    Problem chocoCSP; =0A=
    //............... =0A=
    //-----Set up the problem =0A=
    //............... =0A=
    System.out.println("SolveAll says: " + chocoCSP.solveAll()); =0A=
    System.out.println("isFeasible says: " + chocoCSP.isFeasible()); =0A=
 =0A=
The following was the output =0A=
 =0A=
SolveAll says: true =0A=
isFeasible says: false =0A=
 =0A=
I thought solveAll() returns false if no solution exists. Also, for=0A=
this problem, the solutions I get from the nextSolution() method do not=0A=
satisfy all constraints. Any explainations would be really helpful? =0A=
 =0A=
Thanks, =0A=
-Varun.=0A=
		=09
_________________________________________________________________
Be one of the first to try Windows Live Mail.
http://ideas.live.com/programpage.aspx?versionId=3D5d21c51a-b161-4314-9b0e-=
4911fb2b2e6d=