[lammps-users] group defined by distance from mobile group

Dear all,

I'd like to define a group based on the distance from another group.
That is, group X is a fixed atom or set of atoms in the simulation
(simple based on atom/molecule IDs). Group 2 is the set of atoms
within some distance r from group 1. When dealing with molecules it
might be somewhat more complex given part of a molecule could be in or
out of the region defined by the distance r. In that case there would
have to be a rule defined.

I haven't found a way to do this. Any ideas? I imagine I'll have to
write this up myself but there is a lot of scripting functionality in
LAMMPS with which I'm not too familiar.


There is no command like that in LAMMPS currently.
It would be somewhat difficult to compute in parallel,
since you need to check an atom's position from a list
of other atoms, many of which a processor may not know about.

You do realize that "groups" in LAMMPS are static? Thus
once you define it, and atom's move, then the distance
criterion will no longer hold, but the group assignment
will not change.


I did not realize they were by definition static. This seems like a
nail in the coffin for implementing this in LAMMPS for me. I was
worried about the parallelization, but would have been able to scrape
by with a serial calculation if it was technically feasible.