Prev: Global Variables/Passing Data Using GUIDE
Next: unsupervised image segmentation..need help
From: Hooman on 28 May 2010 13:15
Hello all,

We are in the process of deciding whether we can continue the
development of our simulation models in Simulink. The problem is that
the planned changes and additions to the model will increase the
number of state variables to the order of tens of thousands. Has
anybody on the list simulated a very large system? How large?

I would appreciate any input or comment as how to handle such a case.
We are somehow tied to using Simulink. Within the Simulink framework,
what are the possible solutions?

Thanks,

Hooman
From: Guy Rouleau on 31 May 2010 08:04
Hooman <hooman.javidnia(a)gmail.com> wrote in message <da7e908c-8a3c-44a3-927e-cc1f01c6144c(a)o15g2000vbb.googlegroups.com>...
> Hello all,
>
> We are in the process of deciding whether we can continue the
> development of our simulation models in Simulink. The problem is that
> the planned changes and additions to the model will increase the
> number of state variables to the order of tens of thousands. Has
> anybody on the list simulated a very large system? How large?
>
> I would appreciate any input or comment as how to handle such a case.
> We are somehow tied to using Simulink. Within the Simulink framework,
> what are the possible solutions?
>
> Thanks,
>
> Hooman

Hi,

It is possible for Simulink to handle this type of large model.

To reduce the initialization time, build time and memory required, I recommend dividing your main model using accelerated Model Reference blocks.

Seth introduce this topic here:

http://blogs.mathworks.com/seth/2008/12/05/the-answer-for-large-scale-modeling-model-reference/

You can also find white papers on this topic, for example:

http://www.mathworks.com/mason/tag/proxy.html?dataid=12971&fileid=63037

Guy
From: Hooman on 31 May 2010 11:30
On May 31, 8:04 am, "Guy Rouleau" <guy.rouleau.nos...(a)mathworks.com>
wrote:
> Hooman <hooman.javid...(a)gmail.com> wrote in message <da7e908c-8a3c-44a3-927e-cc1f01c61...(a)o15g2000vbb.googlegroups.com>...
> > Hello all,
>
> > We are in the process of deciding whether we can continue the
> > development of our simulation models in Simulink. The problem is that
> > the planned changes and additions to the model will increase the
> > number of state variables to the order of tens of thousands. Has
> > anybody on the list simulated a very large system? How large?
>
> > I would appreciate any input or comment as how to handle such a case.
> > We are somehow tied to using Simulink. Within the Simulink framework,
> > what are the possible solutions?
>
> > Thanks,
>
> > Hooman
>
> Hi,
>
> It is possible for Simulink to handle this type of large model.
>
> To reduce the initialization time, build time and memory required, I recommend dividing your main model using accelerated Model Reference blocks.
>
> Seth introduce this topic here:
>
> http://blogs.mathworks.com/seth/2008/12/05/the-answer-for-large-scale...
>
> You can also find white papers on this topic, for example:
>
> http://www.mathworks.com/mason/tag/proxy.html?dataid=12971&fileid=63037
>
> Guy

Thanks Guy. I will read Seth's blog post and also the white paper to
see if we can benefit from these features.

Thanks again for the reply and the nice blog posts.

Cheers,
Hooman
 | 
Pages: 1
Prev: Global Variables/Passing Data Using GUIDE
Next: unsupervised image segmentation..need help