Alt_Newton_GD#

class Alt_Newton_GD(*params)[source]#

Bases: _Optimizer

Subclass of Optimizer, inherits its behavior.

Additional Features:#

  • run() is implemented via alternating Newton and gradient descent steps; alternating improves stability

  • _break_condition() is implemented to check for convergence

Methods Summary

run(x0)

Run alternating Newton-GD optimization routine with initial point x0.

Methods Documentation

run(x0)[source]#

Run alternating Newton-GD optimization routine with initial point x0.

Parameters:

x0 (ArrayLike) – Initial point for optimization.

Returns:

  • x_opt (FloatNDArray) – Optimal point found by the optimizer.

  • f_opt (float) – Function value at the optimal point.

  • grad_opt (FloatNDArray) – Gradient of the objective function at the optimal point.

  • hess_opt (FloatNDArray) – Hessian of the objective function at the optimal point.

Return type:

Tuple[ndarray[tuple[int, …], dtype[float64]], float, ndarray[tuple[int, …], dtype[float64]], ndarray[tuple[int, …], dtype[float64]]]

Parameters:

params (Any)