Confidence intervals and different seeds in omnet++ 4.0

[gabry]

Hi all,
I'm pretty new to omnet++ (it's just one month I'm using it), and last
year I used for a short period of time ns-2.
Granted that I searched without success on the online documentation
the solution for my problem, I would like to know how to vary the seed
for different runs of a measurement (i.e. simulation run with the same
parameters), and if/how is possible to set confidence intervals for
them.

Thanks in advance.

Regards,
G. Zannerini.

[gabry]

I found that the .ini option "repeat = X" runs a simulation X times
using different seeds...
Ok, but assuming I want to have a 95% confidence interval on my
results, does exists a way to check it automatically from omnet++?

tia.

br,
GZ.

[Andreas Köpke]

No, you can neither check it from omnet, nor from any other tool currently in
existence. You need to specify the width of your confidence interval, not only
the probability that the searched value lies within.

> --~--~

[Juan-Carlos.M...@sophia.inria.fr]

Hi.


Under the following url I have some scripts to do that. they support
omnet 4.0 vector and scalars files.


there are some explanations how to use them in the same web page.


http://www-sop.inria.fr/members/Juan-Carlos.Maureira_Bravo/OMNeT/Analysis_Scripts


JcM