bug-prolog
[Top][All Lists]

## Re: A question (bug?) on FD solver

 From: Daniel Diaz Subject: Re: A question (bug?) on FD solver Date: Tue, 25 Mar 2003 10:55:32 +0100 User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.2) Gecko/20021120 Netscape/7.01

```See answer I post in users-prolog

Xiaohua Kong wrote:
```
```Hi,
```
I am using FD solver to work on some verification problem that is modeled as CSP problem. Since most constraints are linear, I choose partial consistency operators in FD solver.
```Here is a small example:
```
========================
```ieq(LD1,LD2,Va):-
LD1 = [X1,X2,X3,X4],
LD2 = [D1,D2],
fd_domain(LD1,0,20),
fd_domain(LD2,1,12),
Va #=<5,
Va #>=0,
X1 #= 0,
X2 #=< X1+4,
X2 #>= X1+2,
X3 #= X1+D1,
X4 #= X2+D2,
D1 #=< D2+Va,
D1 #>= D2-Va,
X4 #=< X3.
===========================
LD1 = [0,_#25(2..4),_#47(3..12),_#69(3..12)]
LD2 = [_#91(3..12),_#113(1..10)]
Va = _#134(0..5)
However, there is no assignment exsits if Va=0.
Actually, labeling Va will return give answer that Va could b [2,5].
==============================================
```
Several questions on this penominon: - Is there any problem with the code? Or it is a problem that caused by the algrithm. - Without labeling, if the solver gives answer "yes" and return a domain, does that means there exist at least one solution? (Even though the answer of the variable domains is not accurate).
```- The computation cost of labeling.
```
Thanks
```Xiaohua Kong

--
```
Ce message a subi une analyse antivirus par MailScanner <http://www.mailscanner.info/> ;
```il est vraisemblablement sans danger.

------------------------------------------------------------------------

_______________________________________________
Bug-prolog mailing list
http://mail.gnu.org/mailman/listinfo/bug-prolog
```
```

--
```
Ce message a subi une analyse antivirus par MailScanner ; il est vraisemblablement
```sans danger.

```