basin 0.2.0

An optimization library for Rust
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

#![cfg(feature = "ndarray")]

use basin::problems::Rosenbrock;
use basin::{
    Backtracking, BasicSimplexState, BasicState, CostFunction, Executor, GradientDescent,
    NelderMead,
};
use ndarray::{array, Array1};

#[test]
fn gradient_descent_with_ndarray_array1() {
    let problem = Rosenbrock::<Array1<f64>>::default();
    let initial = array![-1.2, 1.0];
    let initial_cost = problem.cost(&initial);

    let result = Executor::new(
        problem,
        GradientDescent::new(0.001),
        BasicState::new(initial),
    )
    .max_iter(10_000)
    .run();

    assert!(
        result.cost() < initial_cost * 0.1,
        "expected cost to drop by >10x: initial={}, final={}",
        initial_cost,
        result.cost()
    );
}

#[test]
fn gradient_descent_with_ndarray_array1_and_backtracking() {
    let problem = Rosenbrock::<Array1<f64>>::default();
    let initial = array![-1.2, 1.0];
    let initial_cost = problem.cost(&initial);

    let result = Executor::new(
        problem,
        GradientDescent::with_line_search(Backtracking::new()),
        BasicState::new(initial),
    )
    .max_iter(10_000)
    .run();

    assert!(
        result.cost() < initial_cost * 0.1,
        "expected cost to drop by >10x: initial={}, final={}",
        initial_cost,
        result.cost()
    );
}

#[test]
fn nelder_mead_with_ndarray_array1() {
    let problem = Rosenbrock::<Array1<f64>>::default();
    let initial = array![-1.2, 1.0];

    let result = Executor::new(
        problem,
        NelderMead::adaptive(),
        BasicSimplexState::new(initial),
    )
    .max_iter(2_000)
    .run();

    assert!(result.cost() < 1e-6, "cost = {}", result.cost());
}