[gecode-users] Ordering of propagators

[Christian Schulte]

Hi Peter,

yes you are right: propagators with lower cost are most likely to be run
before propagators with higher cost.

The cost values basically capture some levels of constant runtime and then
some levels of asymptotic runtime. The order in which they are defined, also
describes the order in which propagators are likely to be picked for
execution.

The addendum "likely" has to do with the fact that some of the symbolic
costs (for example, PC_BINARY_LO and PC_UNARY_LO) are mapped to the same
actual cost level. For an anlysis of this, please check:
	http://web.it.kth.se/~schulte/paper.html?id=SchulteStuckey:CCoR:2006

What do you mean with how the cost function is used? Could you elaborate,
please?

Cheers
Christian

--
Christian Schulte, http://www.imit.kth.se/~schulte/ 

-----Original Message-----
From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Peter Tiedemann
Sent: Tuesday, April 10, 2007 10:17 AM
To: users at gecode.org
Subject: [gecode-users] Ordering of propagators


Hi,

Is there some way to ensure ( or just strongly suggest) that a propagator is
scheduled to run only after the other propagators have stabilized ( or thats
its just in general pushed to the back of the propagation queue) ?

I thought perhaps the propagation cost() function could achieve an effect
like this, if it returns a very high cost?

Related to this, how is the cost() function used exactly? I cant seem to
find any information on this in the documentation.

Regards,
Peter Tiedemann

_______________________________________________
Gecode users mailing list
users at gecode.org https://www.gecode.org/mailman/listinfo/gecode-users