Re: Molecular Dynamics or Langevin Dynamics

From: Giovanni Bellesia (giovanni.bellesia_at_ucd.ie)
Date: Wed Apr 27 2005 - 19:18:58 CDT

You're right Marc, but this is not exactly what I meant to say.
Once you assume the "ergodicity" of your molecular simulation, Langevin
Dynamics together with Nose-Hoover thermostat (and Nose-Hoover chains)
are all non-hamiltonian schemes for generating the canonical ensemble
(see for example J. Phys. Chem. B 2000, 104, 159-178).
Another common thermostat is Berendsen's. It is quite effective in
generating isothermal conditions, but you cannot prove that the
sampling is a canonical one (deviations are usually quite small
though).
All this in a quick and dirty way from a strict statistical mechanics
point of view.
G.

On 27 Apr 2005, at 18:39, Marc Q. Ma wrote:

> Thanks, G., for your note. It is only true if you simulate your system
> "long enough". An integration scheme does not guarantee simulating the
> right ensemble, it is only providing a basis to achieve the goal.
>
> Marc
> On Apr 27, 2005, at 11:59 AM, Giovanni Bellesia wrote:
>
>>
>>
>> Marc Q. Ma wrote:
>>
>>> you should name it constant temperature molecular dynamics, or
>>> molecular dynamics simulations of canonical ensemble, or MD of
>>> isothermal ensemble.
>>>
>>> Marc
>>> On Apr 26, 2005, at 12:21 PM, sabri bora erdemli wrote:
>>>
>>>> Hi all;
>>>> I am using NAMD for simulating protein complexes. In my simulation
>>>> , I am using Langein Dynamics, which is already implemented in
>>>> NAMD,in order to keep Temperature constant. When I would like to
>>>> name my work, which of the following should I use, Langevin
>>>> Dynamics or Molecular Dynamics?? Let me phrase in anoher way, Am I
>>>> doing Molecular Dynamics or Langevin Dynamics? thanks in
>>>> advance... Sabri Bora Erdemli
>>>> Research and Teaching Asistant
>>>> KOC UNIVERSITY
>>>> Computational Sciences and Engineering
>>>> Koc Universitesi pk.218 34550
>>>> sariyer/Istanbul TURKEY
>>>> tel no: 902123381736
>>>> URL : home.ku.edu.tr/~serdemli
>>>
>>>
>>>
>> A short one just to remind that constant temperature (isothermal) MD
>> does not necessarily mean that you're sampling in the canonical
>> ensemble.
>>
>> G.
>>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:40:42 CST