I am a little confused about the workflow of MD simulation.
The initial velocity of particles in a system is given according to Boltzmann distribution at different temperature.
And then Verlet integration or other intergration methods can be used to get the velocity and position for next time step.
I find that books always talk about potential and force calculation in the workflow of MD simulation. I know the force field to calculate the non-bond and bonded potential. I also know the derivative of potential is force. And many enhanced sampling method can change the force for better sampling. But why the force and potential are needed in the workflow. Verlet integration can calculate the status of next time step. And any modification for better sampling could work directly on velocity and position. The calculation of force and potential seems pointless.
Could anyone give me some suggestions to understand this?
The calculation of the force is not pointless; it's essential. You can't update the positions without the force. F = ma, a = dv/dt. This is covered in the first chapter of any MD textbook (and most simulation software manuals). You can only update the positions of the atoms with the *change* in velocity, so you need the acceleration, hence you need the force.
The calculation of the force is not pointless; it's essential. You can't update the positions without the force. F = ma, a = dv/dt. This is covered in the first chapter of any MD textbook (and most simulation software manuals). You can only update the positions of the atoms with the *change* in velocity, so you need the acceleration, hence you need the force.
From your question I think you mean that it seems pointless to calculate energy and force, but not pointless to calculate only one of them. Is that what you were thinking of?
As Justin described, the force you need since the integrator (Verlet, or other) is based on integrating Newton's second law of motion. Indeed, the energy is not explicitly used to update the positions of the atoms and is thus not strictly needed to perform a simulation. However, the reason they are typically calculated is that they can tell you many interesting things about your system.
verlet can calculate the probabilistic next step.. not the next step in its entirety. Thats why we need force and energy.
For the summary: the knowledge of only particle velocity is not especially useful. Of course one can say where will particle (atom) be in t+1 but the knowledge of force acting on particle is crucial for velocity changes (both speed and direction). Try to imagine some crystalline material in finite temperature, where atoms are maintained in their positions thanks to forces acting on them, when they move out of equilibrium positions.
- Badges
- Science topic