optimization-algorithms

Every optimization algorithm you learned in grad school, in Rust.

What's inside

Algorithm	Module
Gradient descent (SGD, momentum, Adam)	gradient_descent
Conjugate gradient (Fletcher–Reeves, Polak–Ribière)	conjugate_gradient
Quasi-Newton (BFGS, L-BFGS)	quasi_newton
Nelder-Mead simplex	nelder_mead
Simulated annealing	simulated_annealing
Particle swarm optimization	particle_swarm
Line search (backtracking, Wolfe)	line_search
Multi-objective (weighted sum, NSGA-II)	multi_objective
Constraint handling (penalty, barrier)	constraints
Hyperparameter tuning	hyperparameter
Test functions (Rosenbrock, Rastrigin, Ackley, …)	test_functions

Quick start

[dependencies]
optimization-algorithms = "0.1"

use optimization_algorithms::test_functions::{rosenbrock, rosenbrock_gradient};
use optimization_algorithms::quasi_newton::lbfgs;

let result = lbfgs(
    &rosenbrock,
    &rosenbrock_gradient,
    &DVector::from_vec(vec![-1.0, 1.0]),
    &Default::default(),
);
println!("Minimum at {:?}, f = {}", result.x, result.f_x);

License

MIT OR Apache-2.0