object

gradient_descent(Problem)

• Problem - Problem object implementing local_optimization_problem_protocol and defining gradient/2.

Gradient descent (steepest descent) local optimizer with optional backtracking Armijo line search. Requires the problem to define gradient/2. Supports box constraints via projection, minimization and maximization.

Availability:

logtalk_load(local_optimization(loader))

Author: Paulo Moura

Version: 1:0:0

Date: 2026-09-03

Compilation flags:

static, context_switching_calls

Imports:

public local_optimization_solver(Problem)

Uses:

linear_algebra

list

Remarks:

• Line search: The line_search(armijo) option (default) uses backtracking Armijo line search. The line_search(fixed) option uses a constant step size given by step_size(S).

• Bounds: When the problem defines position_bounds/1, the gradient step is projected onto the box (projected gradient descent).

• Direction: For minimization the search direction is -gradient; for maximization it is +gradient.

Inherited public predicates:

 check_option/1  check_options/1  default_option/1  default_options/1  option/2  option/3  run/2  run/3  run/4  valid_option/1  valid_options/1

Public predicates

(no local declarations; see entity ancestors if any)

Protected predicates

(no local declarations; see entity ancestors if any)

Private predicates

(no local declarations; see entity ancestors if any)

Operators

(none)

See also

local_optimization_problem_protocol, local_optimization_solver(Problem), bfgs(Problem), conjugate_gradient(Problem), nelder_mead(Problem)