Iterative Algorithms (sigpy.alg)

This module provides an abstract class Alg for iterative algorithms, and implements commonly used methods, such as gradient methods, Newton’s method, and the augmented Lagrangian method.

The Algorithm Class

sigpy.alg.Alg Abstraction for iterative algorithms.

First-order Gradient Methods

sigpy.alg.GradientMethod First order gradient method.
sigpy.alg.ConjugateGradient Conjugate gradient method.
sigpy.alg.PrimalDualHybridGradient Primal dual hybrid gradient.
sigpy.alg.ADMM Alternating Direction Method of Multipliers.
sigpy.alg.SDMM Simultaneous Direction Method of Multipliers.

Other Methods

sigpy.alg.NewtonsMethod Newton’s Method.
sigpy.alg.PowerMethod Power method to estimate maximum eigenvalue and eigenvector.
sigpy.alg.AltMin Alternating minimization.
sigpy.alg.AugmentedLagrangianMethod Augmented Lagrangian method for constrained optimization.