egobox-doe 0.35.1

A library for design of experiments
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

use crate::SamplingMethod;
use linfa::Float;
use ndarray::{Array, Array1, Array2, ArrayBase, Data, Ix2, s};
use ndarray_stats::QuantileExt;

#[cfg(feature = "serializable")]
use serde::{Deserialize, Serialize};

/// The FullFactorial design consists of all possible combinations
/// of levels for all components within the design space.
#[derive(Clone, Debug)]
#[cfg_attr(feature = "serializable", derive(Serialize, Deserialize))]
pub struct FullFactorial<F: Float> {
    /// Design space definition as
    /// The ith row is the [lower_bound, upper_bound] of xi, the ith component of a sample x
    xlimits: Array2<F>,
}

impl<F: Float> FullFactorial<F> {
    /// Constructor given a design space given a (nx, 2) matrix \[\[lower bound, upper bound\], ...\]
    ///
    /// ```
    /// use egobox_doe::FullFactorial;
    /// use ndarray::arr2;
    ///
    /// let doe = FullFactorial::new(&arr2(&[[0.0, 1.0], [5.0, 10.0]]));
    /// ```
    pub fn new(xlimits: &ArrayBase<impl Data<Elem = F>, Ix2>) -> Self {
        if xlimits.ncols() != 2 {
            panic!("xlimits must have 2 columns (lower, upper)");
        }
        FullFactorial {
            xlimits: xlimits.to_owned(),
        }
    }
}

impl<F: Float> SamplingMethod<F> for FullFactorial<F> {
    fn sampling_space(&self) -> &Array2<F> {
        &self.xlimits
    }

    fn normalized_sample(&self, ns: usize) -> Array2<F> {
        //! the number of level by components is choosen as evenly as possible
        //!
        let nx = self.xlimits.nrows();
        let weights: Array1<F> = Array1::ones(nx) / F::cast(nx);
        let mut num_list: Array1<usize> = Array::ones(nx);

        while num_list.fold(1, |acc, n| acc * n) < ns {
            let w: Array1<F> = &num_list.mapv(|v| F::cast(v)) / F::cast(num_list.sum());
            let ind = (&weights - &w).argmax().unwrap();
            num_list[ind] += 1;
        }
        let nrows = num_list.fold(1, |acc, n| acc * n);
        let mut doe = Array2::<F>::zeros((nrows, nx));

        let mut level_repeat = nrows;
        let mut range_repeat = 1;
        for j in 0..nx {
            let n = num_list[j];
            level_repeat /= n;
            let mut chunk = Array1::zeros(level_repeat * n);
            for i in 0..n {
                let fill = if n > 1 {
                    F::cast(i) / F::cast(n - 1)
                } else {
                    F::cast(i)
                };
                chunk
                    .slice_mut(s![i * level_repeat..(i + 1) * level_repeat])
                    .assign(&Array1::from_elem(level_repeat, fill));
            }
            for k in 0..range_repeat {
                doe.slice_mut(s![n * level_repeat * k..n * level_repeat * (k + 1), j])
                    .assign(&chunk);
            }
            range_repeat *= n;
        }
        doe.slice(s![0..ns, ..]).to_owned()
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use approx::assert_abs_diff_eq;
    use ndarray::{arr2, array};

    #[test]
    fn test_ffact() {
        let xlimits = arr2(&[[5., 10.], [0., 1.]]);
        let expected = array![
            [5., 0.],
            [5., 0.5],
            [5., 1.],
            [7.5, 0.],
            [7.5, 0.5],
            [7.5, 1.],
            [10., 0.],
            [10., 0.5],
            [10., 1.],
        ];
        let actual = FullFactorial::new(&xlimits).sample(9);
        assert_abs_diff_eq!(expected, actual, epsilon = 1e-6);
    }

    #[test]
    fn test_ffact2() {
        let xlimits = arr2(&[
            [-10., 10.],
            [0., 1.],
            [0., 1.],
            [0., 1.],
            [-10., 10.],
            [1., 8.],
        ]);
        let expected = array![
            [-10.0, 0.0, 0.0, 0.0, -10.0, 1.0],
            [-10.0, 0.0, 0.0, 1.0, -10.0, 1.0],
            [-10.0, 0.0, 1.0, 0.0, -10.0, 1.0],
            [-10.0, 0.0, 1.0, 1.0, -10.0, 1.0],
            [-10.0, 1.0, 0.0, 0.0, -10.0, 1.0],
            [-10.0, 1.0, 0.0, 1.0, -10.0, 1.0],
            [-10.0, 1.0, 1.0, 0.0, -10.0, 1.0],
            [-10.0, 1.0, 1.0, 1.0, -10.0, 1.0],
            [10.0, 0.0, 0.0, 0.0, -10.0, 1.0],
            [10.0, 0.0, 0.0, 1.0, -10.0, 1.0]
        ];

        let actual = FullFactorial::new(&xlimits).sample(10);
        assert_abs_diff_eq!(expected, actual, epsilon = 1e-6);
    }
}