Objet : Help-mcsim digest, Vol 1 #134 - 1 msg

[Frédéric BOIS]

Dear Yanyuan

You want to integrate over x and t, right ? Could cast your problem as a system 
of partial differential equations in t
and x. If yes, then you could use a line method to solve the PDEs. MCSim can 
solve well behaved PDE systems
(see the version 5 beta pde exemples, in the tar.gz file attached). If you need 
more sophisticated integration you
may want to try doing it throught a statistical model and MCMC simulations. You 
may need to code some bits by hand
though.

Frederic
 


---- original message -----
Date: Tue, 15 Oct 2002 12:54:17 -0400
From: Yanyuan Ma <address@hidden>
Organization: NC State University
To: address@hidden, address@hidden 
Subject: Is MCSim capable to do this?

Dear Fredric,

   This is a very simle question and I hope you can at least give me a
quick reply. 

   I know that MCSim is designed to deal with problems where the mean
function is implicitly given through an ordinary differential equation
of t.  

   Now, my mean function has one more complexity on top of that. It is
an integral of f(t,x)*g(x). Where g is an explicit function and f is the
solution of some ordinary differential equation. Can MCSim deal this
kind of model?

   You can probably already guess that we are trying to integrate out x
which represents the individual parameters. Indeed this is the case. And
we have to do this because we don't get data of each individual, we are
only given their average data. 
   
Thank you very much in advance

Yanyuan

mcsim-v5beta.tar.gz
Description: Binary data