[gecode-users] argmin

[Christian Schulte]

Hi,

I just made all changes to the trunk (better documentation, pointers into
the GCCAT, offset variants).

Then 1) becomes obsolete. 2) yes, that's the case.

Yep, that is correct.

Cheers
Christian

--
Christian Schulte, www.gecode.org/~schulte
Professor of Computer Science, KTH, cschulte at kth.se
Expert Researcher, SICS, cschulte at sics.se


-----Original Message-----
From: Kish Shen [mailto:kisshen at cisco.com] 
Sent: Friday, June 26, 2015 4:49 AM
To: cschulte at kth.se; users at gecode.org
Subject: Re: [gecode-users] argmin

Hi Christian,

Thanks again.

I have implemented the argmin/max constraints for gfd (the ECLiPSe Gecode
interface), but I have several questions:

1) I am using Ints::Limits::max for the dummy 0th element of the
IntVarArray. Is this the correct choice, or does using such a large number
affect the comlexity of the computation?

2) The MPG says that if tiebreak is false, then the constraint is domain
consistent. However, ICL_DOM is not mentioned in the reference
documentation. Is this because the constraint is not domain consistent for
tiebreak = true?

I have implemented the tiebreak = true and false as 2 constraints, and for
the false case, the constraint can be posted as domain consistent. I hope
this is correct?

Thanks and cheers,

Kish

On 25/06/2015 10:00, Christian Schulte wrote:
> Hi Kish,
>
> Thanks for the points. I'll address all of them a little later (end of 
> semester rush here).
>
> Cheers
> Christian
>
> --
> Christian Schulte, www.gecode.org/~schulte Professor of Computer 
> Science, KTH, cschulte at kth.se Expert Researcher, SICS, 
> cschulte at sics.se
>
> -----Original Message-----
> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On 
> Behalf Of Kish Shen
> Sent: Wednesday, June 24, 2015 10:50 PM
> To: users at gecode.org
> Subject: Re: [gecode-users] argmin
>
>
> Hi Christian,
>
> Thanks for the pointer. I assume as applied here, it is the index for 
> the
> variable(s) in x that has the minimum value.
>
> I did not know (or at least remember) about argmin, and I don't know 
> if it is common knowledge, so perhaps a little more explanation in the 
> MPG might be helpful?
>
> I assume there is no offset that can be specified for this constraint, 
> to offset the index, as is available for some of the other constraints?
> That is, in order to get this constraint so that the index start from 
> 1, I need to add a dummy 0'th element to x?
>
> I think argmin is min_index in GCAT? This is not mentioned in the MPG, 
> as is done for other constraints in the MPG, or am I missing some 
> differences between argmin and min_index? (GCAT's minimum is also not 
> mentioned for min)
>
> Cheers,
>
> Kish
>
> On 24/06/2015 10:45, Christian Schulte wrote:
>> Well it is just argmin:
>> 	https://en.wikipedia.org/wiki/Arg_max
>>
>> Christian
>>
>> --
>> Christian Schulte, www.gecode.org/~schulte Professor of Computer 
>> Science, KTH, cschulte at kth.se Expert Researcher, SICS, 
>> cschulte at sics.se
>>
>>
>> -----Original Message-----
>> From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On 
>> Behalf Of Kish Shen
>> Sent: Wednesday, June 24, 2015 9:41 AM
>> To: users at gecode.org
>> Subject: [gecode-users] argmin
>>
>> Hi,
>>
>> I am trying to understand the argmin/max constraints, but I don't 
>> understand the description in the MPG (section 4.4.5 Arithemetic 
>> constraints),L what does Chargmin(x) mean in:
>>
>> argmin(home, x, y)
>>
>> constrains y to argimin(x)?
>>
>>
>> Cheers,
>>
>> Kish
>>
>>
>> _______________________________________________
>> 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
>