|
|
|
|
Subject: RE : fixed random seed
From: christophe LE PAGE (le_page@cirad.fr)
Date: Tue Jun 24 2008 - 19:12:21 CEST
Bonjour Paul,
In theory you are right!
In practice, if you are using some instances of Set in your model, this may
disturb the random generator…
One way to fix this problem is to use the class IndexedSet instead of the
class Set.
Clp
-----Message d'origine-----
De : owner-cormas@cirad.cirad.fr [mailto:owner-cormas@cirad.cirad.fr] De la
part de Paul Caplat
Envoyé : jeudi 19 juin 2008 15:26
À : Cormas
Objet : fixed random seed
Hi all,
as far as I understand, when the random seed is constant across replicates,
the outcomes should be absolutely similar ?
it does not work in my model...there is variation between replicates as
strong as with a non-fixed seed..Could someone give me insights on this
issue ?
Paul
-- ,,, (o o) --oOO--(_)--OOo----------------------------Dr Paul Caplat postdoctoral fellow Departments of Environmental Biology / Maths & Stats University of Guelph Guelph On, N1G2W1 Canada http://www.uoguelph.ca/~pcaplat
.oooO Oooo. ===( )==( )=========== \ ( ) / \_) (_/