[gecode-users] mod constraint branching

[Guido Tack]

On 14 May 2011, at 21:39, Alin wrote:

> Hallo,
> 
> I have been trying the mod constraint, but I encountered kind of a problem with
> the branching:
> 
> [details about what I did]
> I have resources that can do a job only on a certain day
> so i set the time to be the same as the resource available that day
> 
> mod(*this, resources[i], iDays, start_times[i]);
> 
> [actual problem]
> my problem is that when doing the branching
> if I do:
> 
> branch(*this, resources, INT_VAR_NONE, INT_VAL_RND);
> 
> I get correct values for resources, but 
> 
> branch(*this, resources, INT_VAR_NONE, INT_VAL_MIN);
> 
> will return incorrect values,
> for example: resource[i] will be set to 0,
>             and start_times[i] will be 2
> 
>             while iDay is constant (7)
> 
> why does this happen ?
> shouldn't branching determine just the order of the elements that are searched ?
> shouldn't the value be correct indifferent of the branching type ?

Yes, the branching just determines the order, and the solution should be correct.  I couldn't easily reproduce the problem, but it sounds like a bug in the mod propagator (the VAL_MIN branching just exhibits the bug because it actually tries 0 for resource[i], whereas the random branching might not).
As I said, I can't reproduce this.  Can you send me a short version of your code that exhibits this problem?

Cheers,
	Guido

-- 
Guido Tack, http://people.cs.kuleuven.be/~guido.tack/