[gecode-users] Activity-based search and nondeterministic behaviour

[Christian Schulte]

Hi Andre,

 

Right, they are deterministic, provided:

- you do not use parallel search (as parallel search is indeterministic
itself)

- you keep the same values for recomputation

 

Best

Christian

 

--

Christian Schulte, Professor of Computer Science, KTH,
www.ict.kth.se/~cschulte/

 

From: andre.cire at gmail.com [mailto:andre.cire at gmail.com] On Behalf Of Andre
Augusto Cire
Sent: Friday, February 14, 2014 7:05 PM
To: cschulte at kth.se
Cc: users at gecode.org
Subject: Re: [gecode-users] Activity-based search and nondeterministic
behaviour

 

Hi Christian,

 

Thanks a lot for your answer. Just to clear out any confusion: this means
that, in the latest version (4.2.1), all search types (including AFC) are
deterministic as long as there are no non-deterministic propagators, right?

 

Cheers,

Andre

 

On Mon, Feb 10, 2014 at 3:23 PM, Christian Schulte <cschulte at kth.se> wrote:

Dear Andre,

 

Now it is not deterministic when non-detrministic propagators are around.

 

Moreover, changing the recomputation distance will also (slightly) change
the activity information recorded.

 

Cheers

Christian

 

--

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

 

From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Andre Augusto Cire
Sent: Monday, February 10, 2014 6:43 PM
To: users at gecode.org
Subject: [gecode-users] Activity-based search and nondeterministic behaviour

 

Dear all,

 

I wonder if you could kindly help me with a very quick question. From a
previous forum post and the manual, we know that nonmonotonic propagators
and AFC-based search may result in a nondeterministic behaviour (post here:
https://www.mail-archive.com/users@gecode.org/msg00916.html). 

 

Just to confirm: is activity-based search deterministic?

 

Thank you very much,

Andre

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gecode.org/pipermail/users/attachments/20140217/0a96eb10/attachment.html>