[Pw_forum] relaxation: mistake in finding the lowest energy structure

[Stefano de Gironcoli]

Finding the global minimum of a system may be difficult. Relaxation converge to the nearest local minimum. Chemical intuition may help to make good guesses but sometime it is not enough.

Simulated annealing would be a general purpose very expensive often overkilling never guaranteed approach 

Random sampling, Genetic or evolutionary algorithms are other approaches used 

stefano 
(sent from my phone)

> On 01 Jul 2016, at 17:13, ZARA NBSH <z.nrbsh at gmail.com> wrote:
> 
> Dear all,
> I would like to study the interaction between 2 organic molecules (contains O, H, and C atoms).
> As the first step I relax​ed​ the structure​ ​(​using​ calculation = 'relax' in &control).
> ​M​y problem is that after finishing the run, it could not find the lowest energy s​tructure.
> And I could​ find some new structures (by hand) with​ lower ​energies.
> The ​energy ​difference​,​ ~10​0​ meV​,​ is ​strongly higher​ than the accuracy of energy​ in my calculations​; and also the structures are very different. 
> 
> In my calculation, the​ vacuum​ is large enough​ and forc_conv_thr is 1.0D-5.
> 
> My question is how can I find the lowest energy ​structure?
> 
> Bests,
> ​zara
> ​
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160701/01601e91/attachment.html>