Class EnergyMinimizer

java.lang.Object

org.episteme.natural.chemistry.computational.EnergyMinimizer

public class EnergyMinimizer
extends Object

Energy minimization algorithms. Used to find stable molecular geometries.

Since:

1.0

Author:

Silvere Martin-Michiellot, Gemini AI (Google DeepMind)

Constructor Summary

Constructors

Constructor	Description
EnergyMinimizer()

Method Summary

Modifier and Type	Method	Description
static double	conjugateGradient(Molecule molecule, int maxIterations)	Conjugate Gradient minimization.
static double	steepestDescent(Molecule molecule, int maxIterations, double learningRate, double tolerance)	Steepest Descent minimization.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

EnergyMinimizer

public EnergyMinimizer()

Method Details

steepestDescent

public static double steepestDescent(Molecule molecule,
 int maxIterations,
 double learningRate,
 double tolerance)

Steepest Descent minimization. Iteratively moves atoms against the gradient of potential energy.

Algorithm:

1. Calculate forces F = -grad(E)
2. Update positions: x += learningRate * F
3. Repeat until energy converges

Parameters:

molecule - The molecule to minimize

maxIterations - Maximum steps

learningRate - Step size

tolerance - Energy change tolerance for convergence

Returns:

Final potential energy

conjugateGradient

public static double conjugateGradient(Molecule molecule,
 int maxIterations)

Conjugate Gradient minimization. Faster convergence than steepest descent near minimum. Uses Fletcher-Reeves formula for beta calculation.

Algorithm:

1. d = -grad(E)
2. alpha = line_search(d)
3. x += alpha * d
4. beta = |new_grad|^2 / |old_grad|^2
5. d = -new_grad + beta * d