Re: simulation problem or bug
Bob, Tom:
Thanks for sharing this explanation.
Users might actually be interested in conducting simulations from a model that
terminates with rounding errors for diagnostic or other purposes. I wonder if
the NONMEM project group might consider a different default MSF behavior?
Instead of an error message and exit, perhaps a warning message with
simulations allowed to proceed would be more useful. Of course we can always
run the simulation by manually setting parameters, as Leonid had done.
Thanks,
Marc
Marc R. Gastonguay, Ph.D. ([email protected])
Scientific Director, Metrum Institute (metruminstitute.org)
-- Metrum Institute is a 501(c)3 non-profit organization --
Quoted reply history
On Wednesday, May 25, 2011 at 10:47 PM, Bauer, Robert wrote:
>
> On 5/25/2011 9:43 PM, Bauer, Robert wrote:
> > Leonid:
> > Tom reminded me of our solution, shown below, which you have indeed
> > experienced with the more verbose error message in nm7.2. NONMEM's
> > classic design regarding MSF is that everything from the analysis is
> > captured has to be just right to be able to carry on with it in
> another
> > analysis. Stuart Beal was very conservative about these things. Over
> > time we have to go back and determine which errors from MSF can
> actually
> > be ignored so the user may continue to use the result.
> >
> >
> > Robert J. Bauer, Ph.D.
> > Vice President, Pharmacometrics
> > ICON Development Solutions
> > Tel: (215) 616-6428
> > Mob: (925) 286-0769
> > Email: [email protected]
> > Web: www.icondevsolutions.com
> >
> >
> > -----Original Message-----
> > From: Ludden, Thomas
> > Sent: Wednesday, May 25, 2011 12:46 PM
> > To: Bauer, Robert
> > Subject: FW: [NMusers] simulation problem or bug
> >
> > Bob:
> >
> > Leonid's most recent e-mail reminded me that the "fix" for 7.2 was to
> > give a more informative error message. It was too late in development
> > to make code changes. In Leonid's case I think the rounding error is
> > the real problem.
> >
> > Tom
> >
> > -----Original Message-----
> > From: Ludden, Thomas
> > Sent: Wednesday, May 25, 2011 6:26 AM
> > To: Bauer, Robert
> > Subject: FW: [NMusers] simulation problem or bug
> >
> > Bob:
> >
> > I should have added that I think this has been fixed in 7.2 and
> possibly
> > 7.1.2 as you suggest.
> >
> > Tom
> >
> > -----Original Message-----
> > From: [email protected]
> [mailto:[email protected]]
> > On Behalf Of Bauer, Robert
> > Sent: Tuesday, May 24, 2011 9:59 PM
> > To: Leonid Gibiansky; Fidler,Matt,FORT WORTH,R&D
> > Cc: nmusers
> > Subject: RE: [NMusers] simulation problem or bug
> >
> > Leonid:
> > Could you please try this problem also with NONMEM 7.1.2 and nonmem
> > 7.2.0
> >
> >
> > Robert J. Bauer, Ph.D.
> > Vice President, Pharmacometrics
> > ICON Development Solutions
> > Tel: (215) 616-6428
> > Mob: (925) 286-0769
> > Email: [email protected]
> > Web: www.icondevsolutions.com
> >
> >
> > -----Original Message-----
> > From: [email protected]
> [mailto:[email protected]]
> > On Behalf Of Leonid Gibiansky
> > Sent: Tuesday, May 24, 2011 9:35 AM
> > To: Fidler,Matt,FORT WORTH,R&D
> > Cc: nmusers
> > Subject: Re: [NMusers] simulation problem or bug
> >
> > Matt,
> > I was able to specify parameters explicitly and run simulations. So
> my
> >
> > immediate problem is solved. I am still unsure what is the problem
> with
> > the MSF run.
> > Thanks
> > Leonid
> >
> > --------------------------------------
> > Leonid Gibiansky, Ph.D.
> > President, QuantPharm LLC
> > web: www.quantpharm.com
> > e-mail: LGibiansky at quantpharm.com
> > tel: (301) 767 5566
> >
> >
> >
> >
> > On 5/24/2011 9:14 AM, Fidler,Matt,FORT WORTH,R&D wrote:
> > > Leonid,
> > >
> > > I am unsure why this occurs. I would try the following though:
> > >
> > > - Change the seed number
> > > - Rerun another control stream linked to the MSF output by the
> model
> > and change the SIGDIG to match the number of significant digits
> obtained
> > from your model and run MAXEVAL=0 to create a MSF to run the
> simulation
> > from. (If you run your simulation in this manner)
> > > - Specify the parameters explicitly, and then run the simulation.
> > > - Run the simulation in another version of NONMEM
> > >
> > > Otherwise I'm out of ideas.
> > >
> > > Matt.
> > >
> > > -----Original Message-----
> > > From: [email protected]
> > [mailto:[email protected]] On Behalf Of Leonid Gibiansky
> > > Sent: Monday, May 23, 2011 3:27 PM
> > > To: nmusers
> > > Subject: [NMusers] simulation problem or bug
> > >
> > > Dear All,
> > > I am trying to run simulations starting from the msf file (nm 7.1).
> > > I've done it multiple times for the current project without any
> > > problems. BUT for one specific problem, I see a very strange
> behavior:
> > > The simulation script runs just fine with SUBPROBLEMS=0 option, but
> > > fails with any other value (including 1) with the following error:
> > >
> > > 0MULTIPLE SUB_PROBLEMS IS INCOMPATIBLE WITH OTHER INFORMATION
> > > IN THE CONTROL STREAM AND INPUT MSF
> > >
> > > I am not sure whether it is important but the original run completed
> > > with 134 (rounding error) estimation error and successful covariance
> > > step (with UNCONDITIONAL option)
> > >
> > > Any ideas of what is the problem?
> > >
> > > Thanks!
> > > Leonid
> > >
> > > --
> > > --------------------------------------
> > > Leonid Gibiansky, Ph.D.
> > > President, QuantPharm LLC
> > > web: www.quantpharm.com
> > > e-mail: LGibiansky at quantpharm.com
> > > tel: (301) 767 5566
> > >
> > >
> > >
> > >
> > > This e-mail (including any attachments) is confidential and may be
> > legally privileged. If you are not an intended recipient or an
> > authorized representative of an intended recipient, you are prohibited
> > from using, copying or distributing the information in this e-mail or
> > its attachments. If you have received this e-mail in error, please
> > notify the sender immediately by return e-mail and delete all copies
> of
> > this message and any attachments.
> > >
> > > Thank you.
> >