|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare (strict_types=1); |
|
4
|
|
|
|
|
5
|
|
|
namespace Np\linAlgb\decompositions; |
|
6
|
|
|
|
|
7
|
|
|
use Np\matrix; |
|
8
|
|
|
use Np\vector; |
|
9
|
|
|
use Np\core\lapack; |
|
10
|
|
|
use Np\exceptions\invalidArgumentException; |
|
11
|
|
|
|
|
12
|
|
|
/** |
|
13
|
|
|
* LU |
|
14
|
|
|
* |
|
15
|
|
|
* The LU decomposition is a factorization of a Matrix as the product of a |
|
16
|
|
|
* lower and upper triangular matrix as well as a permutation matrix. |
|
17
|
|
|
* |
|
18
|
|
|
* @package Np |
|
19
|
|
|
* @category Scientific Library |
|
20
|
|
|
* @author ghost (Shubham Chaudhary) |
|
21
|
|
|
* @email [email protected] |
|
22
|
|
|
* @copyright (c) 2020-2021, Shubham Chaudhary |
|
23
|
|
|
*/ |
|
24
|
|
|
class lu { |
|
25
|
|
|
|
|
26
|
|
|
/** |
|
27
|
|
|
* |
|
28
|
|
|
* @param matrix $m |
|
29
|
|
|
* @return self |
|
30
|
|
|
* @throws InvalidArgumentException |
|
31
|
|
|
*/ |
|
32
|
|
|
public static function factory(matrix $m): self { |
|
33
|
|
|
if (!$m->isSquare()) { |
|
34
|
|
|
throw new invalidArgumentException('Matrix must be given.'); |
|
35
|
|
|
} |
|
36
|
|
|
$ipiv = vector::factory($m->col, vector::INT); |
|
37
|
|
|
$ar = $m->copy(); |
|
38
|
|
|
$lp = lapack::getrf($ar, $ipiv); |
|
39
|
|
|
if ($lp != 0) { |
|
40
|
|
|
return null; |
|
41
|
|
|
} |
|
42
|
|
|
$l = matrix::factory($m->col, $m->col); |
|
43
|
|
|
$u = matrix::factory($m->col, $m->col); |
|
44
|
|
|
$p = matrix::factory($m->col, $m->col); |
|
45
|
|
|
for ($i = 0; $i < $m->col; ++$i) { |
|
46
|
|
|
for ($j = 0; $j < $i; ++$j) { |
|
47
|
|
|
$l->data[$i * $m->col + $j] = $ar->data[$i * $m->col + $j]; |
|
48
|
|
|
} |
|
49
|
|
|
$l->data[$i * $m->col + $i] = 1.0; |
|
50
|
|
|
for ($j = $i + 1; $j < $m->col; ++$j) { |
|
51
|
|
|
$l->data[$i * $m->col + $j] = 0.0; |
|
52
|
|
|
} |
|
53
|
|
|
} |
|
54
|
|
|
for ($i = 0; $i < $m->col; ++$i) { |
|
55
|
|
|
for ($j = 0; $j < $i; ++$j) { |
|
56
|
|
|
$u->data[$i * $m->col + $j] = 0.0; |
|
57
|
|
|
} |
|
58
|
|
|
for ($j = $i; $j < $m->col; ++$j) { |
|
59
|
|
|
$u->data[$i * $m->col + $j] = $ar->data[$i * $m->col + $j]; |
|
60
|
|
|
} |
|
61
|
|
|
} |
|
62
|
|
|
for ($i = 0; $i < $m->col; ++$i) { |
|
63
|
|
|
for ($j = 0; $j < $m->col; ++$j) { |
|
64
|
|
|
if ($j == $ipiv->data[$i] - 1) { |
|
65
|
|
|
$p->data[$i * $m->col + $j] = 1; |
|
66
|
|
|
} else { |
|
67
|
|
|
$p->data[$i * $m->col + $j] = 0; |
|
68
|
|
|
} |
|
69
|
|
|
} |
|
70
|
|
|
} |
|
71
|
|
|
unset($ar); |
|
72
|
|
|
unset($ipiv); |
|
73
|
|
|
return new self($l, $u, $p); |
|
74
|
|
|
} |
|
75
|
|
|
|
|
76
|
|
|
/** |
|
77
|
|
|
* |
|
78
|
|
|
* @param matrix $l |
|
79
|
|
|
* @param matrix $u |
|
80
|
|
|
* @param matrix $p |
|
81
|
|
|
*/ |
|
82
|
|
|
protected function __construct(protected matrix $l, protected matrix $u, protected matrix $p) { |
|
83
|
|
|
|
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
/** |
|
87
|
|
|
* Return the lower triangular matrix. |
|
88
|
|
|
* |
|
89
|
|
|
* @return matrix |
|
90
|
|
|
*/ |
|
91
|
|
|
public function l(): matrix { |
|
92
|
|
|
return $this->l; |
|
93
|
|
|
} |
|
94
|
|
|
|
|
95
|
|
|
/** |
|
96
|
|
|
* Return the upper triangular matrix. |
|
97
|
|
|
* |
|
98
|
|
|
* @return matrix |
|
99
|
|
|
*/ |
|
100
|
|
|
public function u(): matrix { |
|
101
|
|
|
return $this->u; |
|
102
|
|
|
} |
|
103
|
|
|
|
|
104
|
|
|
/** |
|
105
|
|
|
* Return the permutation matrix. |
|
106
|
|
|
* |
|
107
|
|
|
* @return matrix |
|
108
|
|
|
*/ |
|
109
|
|
|
public function p(): matrix { |
|
110
|
|
|
return $this->p; |
|
111
|
|
|
} |
|
112
|
|
|
|
|
113
|
|
|
} |
|
114
|
|
|
|
ghostjat /
Np
