Basin Cuts: LJ Clusters

Praharsh Suryadevara

Method

  1. Use CVODE to solve the steepest descent equations 
  2. CVODE is much faster than Gradient descent
  3. CVODE is much faster than scipy implemented stiff ODE methods
  4. Identify minima/optimize with pele (https://github.com/martiniani-lab/pele). Hard fork of Wales group code with additional functionality implemented
  5. Code implemented in basinerror (https://github.com/martiniani-lab/basinerror)

Accuracy curves: Harmonic potential (2d)

Accuracy curves: Harmonic potential (2d)

Random cut around the "global maximum"

\( 8 \times \)equilibrum pairwise distance

Random cut along the Global minimum LJ13

Random cut around the Global minimum LJ13

\( 8 \times \)equilibrum pairwise distance

\( 8 \times \)equilibrum pairwise distance

Colors not matched!!

Random cut around a local minimum

Random cut around a local minimum LJ 13

\( 8 \times \)equilibrum pairwise distance

\( 8 \times \)equilibrum pairwise distance

Colors not matched!!