Re: minimization of the squared gradient

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Tue May 20 2008 - 08:19:58 CDT

Hi Krishnan,
you can find the relevant functions for the conjugate gradient
minimization that namd uses in Controller.C and Sequencer.C. However, to
act to minimize a different quantity you'd need to alter the results
being fed to these routines by the compute objects. As Victor noted,
this is complicated by the fact that namd doesn't have built in
facilities for calculating second derivatives (you could in principle
write some tcl code to do this numerically by perturbing the
coordinates, but this would not be a fun exercise). For saddle point
searching you might be better off using a different package.

On a related note, I'm not sure what kind of system you're planning to
do this on, but please keep in mind that most large solvated
biomolecules (ie, most of the systems people simulate with namd) have
such complicated potential energy surfaces that there are likely to be
very many critical points, the majority of which are probably not
particularly significant.

Best,
Peter

Krishnan, Marimuthu wrote:
> Hi all,
>
> Please correct me if I am wrong.
>
> NAMD has routines to calculate the gradient of the potential energy (force evaluation routine) and
> for functional minimization. So what is required is to square of the gradient and feed it to
> the functional minimization routine. If would appreciate if NAMD-experts can provide me a good solution
> for this problem.
>
> Thanks
> Krishnan
>
> *************************************************
>
> Dr. Marimuthu Krishnan
> Postdoctoral Research Scholar,
> Center for Molecular Biophysics,
> Oak Ridge National Laboratory,
> PO Box 2008 MS6164,
> Oak Ridge, TN 37831-6164,
> Tel : 001-865-574-6123
> Fax : 001-865-576-7651
> E-mail : krishnanm_at_ornl.gov
>
> *************************************************
>
>
>
> -----Original Message-----
> From: owner-namd-l_at_ks.uiuc.edu on behalf of Peter Freddolino
> Sent: Mon 5/19/2008 10:20 AM
> To: Krishnan, Marimuthu
> Cc: namd-l_at_ks.uiuc.edu
> Subject: Re: namd-l: minimization of the squared gradient
>
> Ah, sorry, misread the question. No, I'm not aware of a way to do this
> in namd; you'd need to write a different minimization routine;.
> Peter
>
> Krishnan, Marimuthu wrote:
>
>> Hi Peter,
>>
>> Yes. The "minimize" command minimizes the potential energy (U) of the system. I would like to minimize the squared gradient ([dU]^2) of the potential energy. This way, the system can be quenched to a stationary point (This can correspond to a minimum or a saddle).
>>
>> Thanks.
>> Krishnan
>>
>> *************************************************
>>
>> Dr. Marimuthu Krishnan
>> Postdoctoral Research Scholar,
>> Center for Molecular Biophysics,
>> Oak Ridge National Laboratory,
>> PO Box 2008 MS6164,
>> Oak Ridge, TN 37831-6164,
>> Tel : 001-865-574-6123
>> Fax : 001-865-576-7651
>> E-mail : krishnanm_at_ornl.gov
>>
>> *************************************************
>>
>> ________________________________
>>
>> From: owner-namd-l_at_ks.uiuc.edu on behalf of Peter Freddolino
>> Sent: Mon 5/19/2008 7:40 AM
>> To: Krishnan, Marimuthu
>> Cc: namd-l_at_ks.uiuc.edu
>> Subject: Re: namd-l: minimization of the squared gradient
>>
>>
>>
>> Hi Krishnan,
>> it isn't exactly clear to me from what you wrote: Do you want something
>> different from what you get by using the "minimize" command?
>> Best,
>> Peter
>>
>> Krishnan, Marimuthu wrote:
>>
>>
>>> Dear all,
>>>
>>> Is it possible to use NAMD to minimize the squared gradient of the potential energy? I would appreciate If someone could provide me a script for doing this.
>>>
>>> Thanks.
>>> Krishnan
>>>
>>> *************************************************
>>>
>>> Dr. Marimuthu Krishnan
>>> Postdoctoral Research Scholar,
>>> Center for Molecular Biophysics,
>>> Oak Ridge National Laboratory,
>>> PO Box 2008 MS6164,
>>> Oak Ridge, TN 37831-6164,
>>> Tel : 001-865-574-6123
>>> Fax : 001-865-576-7651
>>> E-mail : krishnanm_at_ornl.gov
>>>
>>> *************************************************
>>>
>>>
>>>
>>
>>
>>
>
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:28 CST