Archived I'm a computational scientist, ask me stuff about computer simulations! (whatever)
submitted ago by The_Cat
Posted by: The_Cat
Posting time: 3.6 years ago on
Last edit time: never edited.
Archived on: 7/28/2017 10:00:00 AM
Views: 519
SCP: 17
18 upvotes, 1 downvotes (95% upvoted it)
Archived I'm a computational scientist, ask me stuff about computer simulations! (whatever)
submitted ago by The_Cat
view the rest of the comments →
[–] Fibbideh 0 points 1 point 1 point (+1|-0) ago
What is your favorite research project you have participated in?
Did you ever have a project that made you feel like you were in way over your head?
If you can, try not to simplify it or dumb it down for me to understand. I know I won't but I get a kick out of trying anyway.
[–] The_Cat [S] 0 points 1 point 1 point (+1|-0) ago
Funny enough, my answer to both those questions is the same: trying to implement a thermostat for a fermionic molecular dynamics simulation.
Molecular dynamics just means that you start from the laws of mechanics/dynamics, start with an initial configuration and follow the system through time. A thermostat keeps the simulation at a constant temperature.
Now, fermionic molecular dynamics (FMD) is a method to simulate a system of indistinguishable particles, fermions, like electrons or neutrons for example. The clever thing about it is that instead of going full quantum, these particles are considered as semiclassical wave packet, a 3D gaussian distribution that roughly keeps it shape.
Fermions are antisymmetric upon exchange, meaning that you cannot have two in the same location, and they will repel each other when they get close. This interaction you can encode by considering it as a curvature of state space (the space of all positions and momenta). So the system follows a semiclassical path through curved state space, where the size of state space is significantly contracted for points when two fermions are close together in position and momentum. This causes the fermions to naturally avoid each other, and is an exact implementation of the Pauli repulsion within a semiclassical context.
Happy googling :p