optimization.optimizer.scipy.orthogonal_descent

Orthogonal Descent Optimizer Module

This module contains the OrthogonalDescent class, which implements a coordinate descent optimization algorithm. The algorithm sequentially optimizes each variable while holding others fixed.

Kramer Harrison, 2026

Classes

OrthogonalDescent(problem)

Orthogonal Descent (Coordinate Descent) optimizer.
class OrthogonalDescent(problem: OptimizationProblem)[source]

Orthogonal Descent (Coordinate Descent) optimizer.

This optimizer minimizes the objective function by sequentially optimizing each variable one at a time. It is useful when derivatives are not available or unreliable.

optimize(max_iter=100, tol=0.0001, plot=False)[source]

Run the orthogonal descent optimization.

Parameters:

  • max_iter (int) – Maximum number of full cycles through all variables.

  • tol (float) – Tolerance for convergence (relative change in cost function).

  • plot – If True, update live plots during optimization.