Developing Potentials

How can I develop potentials for my own work? I’m currently working on high entropy alloys system and need a potential that can describe a system of 4 elements.

Hi
You might use eam/alloy pair_style. There are numerous LAMMPS works with.
Regards
Pascal

How can I develop potentials for my own work? I'm currently working on high entropy alloys system and need a potential that can describe a system of 4 elements.

there are lots of publications describing the parameterization of
potentials. which is most useful depends a lot on the details of your
simulation and the degree of accuracy you are looking for.
for some potentials (e.g. plain EAM potentials or Lennard-Jones)
mixing rules exist, that can be used to automatically derive mixed
terms from single atom type parameters. those are usually of lower
accuracy than parameters specifically derived for a specific mixture.
also you have to keep in mind, that (empirical) potentials are usually
parameterized for a specific thermodynamic state (or range), and that
some functional forms are not flexible to reproduce experimental data
well for certain classes of compounds (e.g. Lennard-Jones vs EAM -> no
embedding term with LJ, or EAM vs. MEAM -> no directional terms with
EAM).
if you look into the tools/eam_database folder, you can find a tool
and reference to a publication to generate specific alloy EAM
potentials for any combination of the supported elements.

there are more complex functional forms (e.g. ReaxFF) that can handle
much more complex interactions and mixtures of elements/compounds, but
those are much more difficult to parameterize (in those cases the
boundaries between science and black art are not so easy to tell
anymore).

at any rate, a thorough study of the published literature of studies
of systems similar to yours is always a good starting point. it will
inform you what people in the field are doing and provide some
assessment of what level of accuracy and what degree of computational
cost you will have to deal with.

axel.