Re: "Crystall Ball" and "@Risk" Users using Gnumeric

[Erich Neuwirth <erich neuwirth univie ac at>]

For Monte Carlo,
the "at most 2 inputs" restruction is can be circimvented.

Example, the Excel way
In A1:A10 put =RAND()
in B1 put =SUM(A1:A10)

in F2:F101 put the numbers 1 to 100
in G1 put =B1


Select F1:G101 and create a data table with
column input cell D1. This is a dummy, nothing is dependent on D1.
Each recalculation step will recompute ALL the random numbers in A1:A10,
therefore you can have as many random inputs in your model.


The trick is to use a dummy input cell just to trigger the generation of
new random numbers.





Gideon Butalid wrote:
> One Gnumeric developer, I think it was Jody Goldberg, wrote that Data
> Tables are supported in Gnumeric though the GUI isn't present yet.
> Additionally Data Tables are only applicable for models if you have two
> uncertain variables. I have more than two most of the time. I think I
> have to use static random numbers, crude old style monte carlo
> simulation. Thanks though, nice link by the way.
> 
> On 7/27/07, *Erich Neuwirth* <erich neuwirth univie ac at
> <mailto:erich neuwirth univie ac at>> wrote:
> 
>     In Excel, you do something like what you describe without addins by
>     using the Data Table menu item.
> 
>     http://ite.pubs.informs.org/Vol1No1/Evans/
> 
>     illustrates the approach
> 
>     Gnumeric does not seem to have Data Table.
>     So this is an item for the wish list ;-)
>     gont to the developers ;-)
> 
> 
> 
> 
> 
> 
>     Gideon Butalid wrote:
>     > I am dual booting Ubuntu and Windows XP in my laptop. As an
>     "analyst" I
>     > build my spreadsheet models (including random numbers) in Gnumeric in
>     > Ubuntu, then I revert back to Windows XP so I could run my model in
>     > Crystal Ball (Excel add-in). I was able to cope in a way that
>     > probability distributions of inputs and their correlations are
>     embedded
>     > in the model--thus I can easily build a spreadsheet model in Gnumeric.
>     > The only use for Crystal Ball is for me to capture the
>     distribution of
>     > the OUTPUT and study how it is distributed.
>     >
>     > This is my question:
>     >
>     > Is it possible to capture the set of outputs in a particular cell in
>     > Gnumeric, and study them later? For instance, a cell will generate
>     a new
>     > output after I press F9 because a new set of random numbers are
>     > generated.  Using the Gnumeric simulation tool will give the mean,
>     > variance etc, but I need the final outputs, that is output per
>     trial, so
>     > I could study how they are distributed.
>     >
>     > Thanks in advance.
>     >
>     > --
>     > gbutalid
>     >
>     >
>     >
>     ------------------------------------------------------------------------
>     >
>     > _______________________________________________
>     > gnumeric-list mailing list
>     > gnumeric-list gnome org <mailto:gnumeric-list gnome org>
>     > http://mail.gnome.org/mailman/listinfo/gnumeric-list
> 
> 
>     --
>     Erich Neuwirth, University of Vienna
>     Faculty of Computer Science
>     Computer Supported Didactics Working Group
>     Visit our SunSITE at http://sunsite.univie.ac.at
>     Phone: +43-1-4277-39464 Fax: +43-1-4277-39459
> 
> 
> 
> 
> -- 
> Gideon Adriel E. Butalid


-- 
Erich Neuwirth, University of Vienna
Faculty of Computer Science
Computer Supported Didactics Working Group
Visit our SunSITE at http://sunsite.univie.ac.at
Phone: +43-1-4277-39464 Fax: +43-1-4277-39459