Re: Solving for the dielectric constant that yields a certain gap width

Hi Ian,

You have the right idea. Rather than writing you own binary search,  
I'd recommend using one of the goal seeking functions such as find- 
root or minimize that are included in libctl. See http://ab-initio.mit.edu/wiki/index.php/Libctl_User_Reference#Miscellaneous_utilities 
  for more information on these functions. I assume you've already  
figured out how to access the bandgap in the ctl file (see http://ab-initio.mit.edu/wiki/index.php/MPB_User_Reference#Output_Variables 
  if not).

I was recently asked to fit some structure parameters to provided  
bandstructure data. I've attached the ctl file I developed in the hope  
that you may find it useful. Since I was fitting to the multiple  
points on the band structure, I used the minimize function to reduce  
the sum of the squares of the difference. I am new to Scheme  
programming so I can't vouch for the quality of the attached script  
other than to say that it worked for me.

-Mark Patterson
Department of Physics
Queen's University at Kingston




On 2008-09-09, at 22:52 , Ian D. Hosein wrote:

> MPB Users & Steven,
>
>    I'm trying to solve a reverse problem of sort. I want set the  
> bandgap of
> my 3D dielectric structure to a certain width, say 1%, and then  
> solve for
> the high index material dielectric value that will yield it, within a
> certain tolerance ofcourse.  For example, an inverted opal would be  
> one of
> my structures.
>    My first guess on how to approach this is to give the the program  
> a min
> and max dielectric value in its search, and I start with the median  
> of the
> two. In a loop, I calculate the band structure and if the gap is  
> greater
> that 1% I need to change the dielectric value to between the min and  
> the
> median, in other words the median becomes the new max. If the gaps  
> is less
> that 1%, then I need to work with a dielectric value between the  
> median and
> the max.  It's sorta like a binary search. I would have to put in some
> conditions so that the program doesn't search forever for a complete  
> gap if
> there isn't ever one.
>    Does this sound right? Anyone have any thoughts on this, I would
> appreciate any suggestions.
>
> Thanks,
>
> Ian
>
>
>
>
> _______________________________________________
> mpb-discuss mailing list
> [email protected]
> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss

_______________________________________________
mpb-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss