Collection of numerical simulations of different physical phenomena
Time-evolution of the 1D Schrodinger equation, where a particle (gaussian wave packet) hits a potential barrier with an energy slightly higher than the particle energy.
Time-evolution of 2D Schrodinger equation, where a 2D gaussian wave packet travels through a double slit. Simulated using alternating direction implicit method over a 1024x1024 grid. For all of the plots, the color represents the phase angle between the real and imaginary parts of the equation.
Visualized in two-dimensions
Visualized in three-dimensions
Demonstrates the use of a one-step residual neural network for data-driven equation approximation. The method aims to be useful for recovering unknown governing equations based only on coarsely distributed observational data of trajectories. It is demonstrated on a dynamical system of a damped pendulum, with a dataset of trajectories generated by an accurate ODE solver. Animation shows the model's predictions on a sample data point at each epoch of training.
Monte Carlo algorithm is used to simulate the equilibrium distribution of the Ising model in two dimensions.
Simulation of particles that interact using a Lennard-Jones potential.