minuit2 0.4.2

Pure Rust port of CERN Minuit2 parameter optimization engine
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
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180

mod common;

use minuit2::{FCN, MnSimplex};

/// Rosenbrock function: f(x,y) = (1-x)^2 + 100(y-x^2)^2
/// Minimum at (1, 1) with f = 0.
#[test]
fn rosenbrock_2d() {
    let result = MnSimplex::new()
        .add("x", 0.0, 0.1)
        .add("y", 0.0, 0.1)
        // With ROOT-compatible Simplex EDM criterion, use a tighter tolerance
        // for Rosenbrock to ensure full convergence.
        .tolerance(1.0e-5)
        .minimize(&|p: &[f64]| (1.0 - p[0]).powi(2) + 100.0 * (p[1] - p[0] * p[0]).powi(2));

    assert!(result.is_valid(), "minimization should converge");

    let params = result.params();
    assert!(
        (params[0] - 1.0).abs() < 0.1,
        "x should be near 1.0, got {}",
        params[0]
    );
    assert!(
        (params[1] - 1.0).abs() < 0.1,
        "y should be near 1.0, got {}",
        params[1]
    );
    assert!(
        result.fval() < 0.01,
        "fval should be near 0, got {}",
        result.fval()
    );
}

/// Simple quadratic bowl: f(x,y) = x^2 + y^2
/// Minimum at (0, 0) with f = 0.
#[test]
fn quadratic_bowl() {
    let result = MnSimplex::new()
        .add("x", 5.0, 1.0)
        .add("y", -3.0, 1.0)
        .minimize(&|p: &[f64]| p[0] * p[0] + p[1] * p[1]);

    assert!(result.is_valid(), "minimization should converge");

    let params = result.params();
    assert!(
        params[0].abs() < 0.01,
        "x should be near 0, got {}",
        params[0]
    );
    assert!(
        params[1].abs() < 0.01,
        "y should be near 0, got {}",
        params[1]
    );
    assert!(
        result.fval() < 1e-4,
        "fval should be near 0, got {}",
        result.fval()
    );
}

/// Gaussian fit to synthetic data.
/// Three parameters: amplitude, mean, sigma.
#[test]
fn gaussian_fit() {
    // Generate synthetic data: Gaussian with amp=10, mean=5, sigma=2
    let true_amp = 10.0;
    let true_mean = 5.0;
    let true_sigma = 2.0;

    let n_data = 50;
    let x_data: Vec<f64> = (0..n_data).map(|i| i as f64 * 0.2).collect();
    let y_data: Vec<f64> = x_data
        .iter()
        .map(|&x| true_amp * (-(x - true_mean).powi(2) / (2.0 * true_sigma * true_sigma)).exp())
        .collect();

    // Chi-square: sum of (model - data)^2
    let chi2 = move |p: &[f64]| -> f64 {
        let amp = p[0];
        let mean = p[1];
        let sigma = p[2];
        x_data
            .iter()
            .zip(y_data.iter())
            .map(|(&x, &y)| {
                let model = amp * (-(x - mean).powi(2) / (2.0 * sigma * sigma)).exp();
                (model - y).powi(2)
            })
            .sum()
    };

    let result = MnSimplex::new()
        .add("amp", 8.0, 1.0)
        .add("mean", 4.0, 0.5)
        .add_lower_limited("sigma", 1.5, 0.5, 0.01)
        .tolerance(0.001)
        .minimize(&chi2);

    assert!(result.is_valid(), "gaussian fit should converge");

    let state = result.user_state();
    let amp = state.value("amp").unwrap();
    let mean = state.value("mean").unwrap();
    let sigma = state.value("sigma").unwrap();

    assert!(
        (amp - true_amp).abs() < 0.5,
        "amp should be near {true_amp}, got {amp}"
    );
    assert!(
        (mean - true_mean).abs() < 0.5,
        "mean should be near {true_mean}, got {mean}"
    );
    assert!(
        (sigma - true_sigma).abs() < 0.5,
        "sigma should be near {true_sigma}, got {sigma}"
    );
}

/// Test bounded parameters with the sin transform.
#[test]
fn bounded_parameters() {
    // Minimize (x-3)^2 with x in [0, 5]
    let result = MnSimplex::new()
        .add_limited("x", 1.0, 0.5, 0.0, 5.0)
        .tolerance(0.001)
        .minimize(&|p: &[f64]| (p[0] - 3.0).powi(2));

    assert!(result.is_valid(), "bounded min should converge");
    let x = result.params()[0];
    assert!((x - 3.0).abs() < 0.1, "x should be near 3.0, got {x}");
    assert!((0.0..=5.0).contains(&x), "x should be within bounds");
}

/// Test with a fixed parameter.
#[test]
fn fixed_parameter() {
    struct QuadWithFixed;
    impl FCN for QuadWithFixed {
        fn value(&self, p: &[f64]) -> f64 {
            (p[0] - 2.0).powi(2) + (p[1] - 3.0).powi(2)
        }
    }

    let result = MnSimplex::new()
        .add("x", 0.0, 0.5)
        .add("y", 0.0, 0.5)
        .fix(1) // fix y at 0
        .minimize(&QuadWithFixed);

    assert!(result.is_valid());
    let params = result.params();
    // x should minimize to 2.0
    assert!(
        (params[0] - 2.0).abs() < 0.1,
        "x should be near 2.0, got {}",
        params[0]
    );
    // y should stay at 0.0 (fixed)
    assert!(
        (params[1] - 0.0).abs() < 1e-15,
        "y should be 0.0 (fixed), got {}",
        params[1]
    );
}

/// Display output should not panic.
#[test]
fn display_output() {
    let result = MnSimplex::new()
        .add("x", 1.0, 0.1)
        .minimize(&|p: &[f64]| p[0] * p[0]);

    common::assert_function_minimum_display(&result, &["x"]);
}