| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | // -*- mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- | ||
| 2 | // vi: set et ts=4 sw=4 sts=4: | ||
| 3 | // | ||
| 4 | // SPDX-FileCopyrightInfo: Copyright © DuMux Project contributors, see AUTHORS.md in root folder | ||
| 5 | // SPDX-License-Identifier: GPL-3.0-or-later | ||
| 6 | // | ||
| 7 | /*! | ||
| 8 | * \file | ||
| 9 | * \ingroup Core | ||
| 10 | * \brief Some tools for random number generation | ||
| 11 | */ | ||
| 12 | #ifndef DUMUX_COMMON_RANDOM_HH | ||
| 13 | #define DUMUX_COMMON_RANDOM_HH | ||
| 14 | |||
| 15 | #include <random> | ||
| 16 | #include <type_traits> | ||
| 17 | #include <cstdint> | ||
| 18 | |||
| 19 | namespace Dumux { | ||
| 20 | |||
| 21 | /*! | ||
| 22 | * \brief A simple uniform distribution | ||
| 23 | * based on a biased uniform number generator | ||
| 24 | * \note Use this if you need a fast library implementation independent generator | ||
| 25 | * without strict requirements about the bias | ||
| 26 | * \note We try to stay close to https://en.cppreference.com/w/cpp/numeric/random/uniform_real_distribution | ||
| 27 | */ | ||
| 28 | template<class Scalar = double> | ||
| 29 | class SimpleUniformDistribution | ||
| 30 | { | ||
| 31 | struct Parameters | ||
| 32 | { | ||
| 33 | Parameters(Scalar a, Scalar b) | ||
| 34 | : a_(a), b_(b) {} | ||
| 35 | |||
| 36 | Scalar a() const { return a_; } | ||
| 37 | Scalar b() const { return b_; } | ||
| 38 | private: | ||
| 39 | Scalar a_, b_; | ||
| 40 | }; | ||
| 41 | public: | ||
| 42 | using param_type = Parameters; | ||
| 43 | using result_type = Scalar; | ||
| 44 | |||
| 45 | ✗ | explicit SimpleUniformDistribution(Scalar min, Scalar max = 1.0) | |
| 46 | : offset_(min) | ||
| 47 |
1/7✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
✗ Branch 2 not taken.
✗ Branch 4 not taken.
✗ Branch 5 not taken.
✗ Branch 7 not taken.
✗ Branch 8 not taken.
|
3 | , size_(max-min) |
| 48 | {} | ||
| 49 | |||
| 50 | explicit SimpleUniformDistribution(const Parameters& p) | ||
| 51 | : SimpleUniformDistribution(p.a(), p.b()) | ||
| 52 | {} | ||
| 53 | |||
| 54 | SimpleUniformDistribution() | ||
| 55 | : SimpleUniformDistribution(0.0) | ||
| 56 | {} | ||
| 57 | |||
| 58 | void param(const Parameters& p) | ||
| 59 | { | ||
| 60 | offset_ = p.a(); | ||
| 61 | size_ = p.b()-p.a(); | ||
| 62 | } | ||
| 63 | |||
| 64 | Parameters param() const | ||
| 65 | { return { offset_, offset_+size_ }; } | ||
| 66 | |||
| 67 | Scalar a() const { return offset_; } | ||
| 68 | Scalar b() const { return offset_ + size_; } | ||
| 69 | |||
| 70 | template<class Generator, | ||
| 71 | typename std::enable_if_t<std::is_same_v<typename Generator::result_type, std::uint_fast32_t>, int> = 0> | ||
| 72 | ✗ | Scalar operator()(Generator& gen) | |
| 73 |
0/7✗ Branch 1 not taken.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
✗ Branch 4 not taken.
✗ Branch 5 not taken.
✗ Branch 8 not taken.
✗ Branch 9 not taken.
|
105 | { return offset_ + size_*(0x1.0p-32 * gen()); } |
| 74 | |||
| 75 | private: | ||
| 76 | Scalar offset_; | ||
| 77 | Scalar size_; | ||
| 78 | }; | ||
| 79 | |||
| 80 | /*! | ||
| 81 | * \brief A simple normal distribution | ||
| 82 | * based on a biased uniform number generator and the Box-Mueller transform | ||
| 83 | * \note Use this if you need a fast library implementation independent generator | ||
| 84 | * without strict requirements about the bias | ||
| 85 | * \note We try to stay close to https://en.cppreference.com/w/cpp/numeric/random/normal_distribution | ||
| 86 | */ | ||
| 87 | template<class Scalar = double> | ||
| 88 | class SimpleNormalDistribution | ||
| 89 | { | ||
| 90 | struct Parameters | ||
| 91 | { | ||
| 92 | Parameters(Scalar m, Scalar s) | ||
| 93 | : m_(m), s_(s) {} | ||
| 94 | |||
| 95 | Scalar m() const { return m_; } | ||
| 96 | Scalar s() const { return s_; } | ||
| 97 | private: | ||
| 98 | Scalar m_, s_; | ||
| 99 | }; | ||
| 100 | public: | ||
| 101 | using param_type = Parameters; | ||
| 102 | using result_type = Scalar; | ||
| 103 | |||
| 104 | 8 | explicit SimpleNormalDistribution(Scalar mean, Scalar stddev = 1.0) | |
| 105 | : mean_(mean) | ||
| 106 | 8 | , stddev_(stddev) | |
| 107 | {} | ||
| 108 | |||
| 109 | explicit SimpleNormalDistribution(const Parameters& p) | ||
| 110 | : SimpleNormalDistribution(p.m(), p.s()) | ||
| 111 | {} | ||
| 112 | |||
| 113 | SimpleNormalDistribution() | ||
| 114 | : SimpleNormalDistribution(0.0) | ||
| 115 | {} | ||
| 116 | |||
| 117 | void param(const Parameters& p) | ||
| 118 | { | ||
| 119 | mean_ = p.m(); | ||
| 120 | stddev_ = p.s(); | ||
| 121 | } | ||
| 122 | |||
| 123 | Parameters param() const | ||
| 124 | { return { mean_, stddev_ }; } | ||
| 125 | |||
| 126 | Scalar m() const { return mean_; } | ||
| 127 | Scalar s() const { return stddev_; } | ||
| 128 | |||
| 129 | template<class Generator> | ||
| 130 | 17989 | Scalar operator()(Generator& gen) | |
| 131 | { | ||
| 132 |
2/2✓ Branch 0 taken 8993 times.
✓ Branch 1 taken 8996 times.
|
17989 | if (isCached_) |
| 133 | { | ||
| 134 | 8993 | isCached_ = false; | |
| 135 | 8993 | return cachedValue_; | |
| 136 | } | ||
| 137 | |||
| 138 | // Box-Mueller transform (https://en.wikipedia.org/wiki/Box%E2%80%93Muller_transform) | ||
| 139 | 8996 | const auto [u1, u2] = generateUniformPair_(gen); | |
| 140 | |||
| 141 | using std::sqrt; using std::log; | ||
| 142 | using std::cos; using std::sin; | ||
| 143 | 8996 | constexpr Scalar twoPi = 2.0 * M_PI; | |
| 144 | |||
| 145 | 8996 | const Scalar magnitude = stddev_ * sqrt(-2.0 * log(u1)); | |
| 146 | 8996 | const Scalar z0 = magnitude * cos(twoPi * u2) + mean_; | |
| 147 | 8996 | const Scalar z1 = magnitude * sin(twoPi * u2) + mean_; | |
| 148 | 8996 | cachedValue_ = z0; | |
| 149 | 8996 | isCached_ = true; | |
| 150 | 8996 | return z1; | |
| 151 | } | ||
| 152 | |||
| 153 | private: | ||
| 154 | template<class Generator, | ||
| 155 | typename std::enable_if_t<std::is_same_v<typename Generator::result_type, std::uint_fast32_t>, int> = 0> | ||
| 156 | ✗ | auto generateUniformPair_(Generator& gen) | |
| 157 | { | ||
| 158 | // biased uniform number generator (0,1) | ||
| 159 | // https://www.pcg-random.org/posts/bounded-rands.html | ||
| 160 | ✗ | constexpr Scalar eps = std::numeric_limits<Scalar>::epsilon(); | |
| 161 | ✗ | Scalar u1 = 0.0, u2 = 0.0; | |
| 162 | do { | ||
| 163 | ✗ | u1 = 0x1.0p-32 * gen(); | |
| 164 | ✗ | u2 = 0x1.0p-32 * gen(); | |
| 165 | ✗ | } while (u1 <= eps); | |
| 166 | ✗ | return std::make_pair(u1, u2); | |
| 167 | ✗ | } | |
| 168 | |||
| 169 | Scalar mean_; | ||
| 170 | Scalar stddev_; | ||
| 171 | bool isCached_ = false; | ||
| 172 | Scalar cachedValue_ = {}; | ||
| 173 | }; | ||
| 174 | |||
| 175 | /*! | ||
| 176 | * \brief A simple log-normal distribution | ||
| 177 | * \note Use this if you need a fast library implementation independent generator | ||
| 178 | * without strict requirements about the bias | ||
| 179 | * \note We try to stay close to https://en.cppreference.com/w/cpp/numeric/random/lognormal_distribution | ||
| 180 | */ | ||
| 181 | template<class Scalar = double> | ||
| 182 | class SimpleLogNormalDistribution | ||
| 183 | { | ||
| 184 | using Parameters = typename SimpleNormalDistribution<Scalar>::param_type; | ||
| 185 | public: | ||
| 186 | using param_type = Parameters; | ||
| 187 | using result_type = Scalar; | ||
| 188 | |||
| 189 | explicit SimpleLogNormalDistribution(Scalar mean, Scalar stddev = 1.0) | ||
| 190 |
3/6✓ Branch 1 taken 5 times.
✗ Branch 2 not taken.
✓ Branch 4 taken 2 times.
✗ Branch 5 not taken.
✓ Branch 7 taken 1 times.
✗ Branch 8 not taken.
|
8 | : normal_(mean, stddev) |
| 191 | {} | ||
| 192 | |||
| 193 | explicit SimpleLogNormalDistribution(const Parameters& p) | ||
| 194 | : SimpleLogNormalDistribution(p.mean, p.stddev) | ||
| 195 | {} | ||
| 196 | |||
| 197 | SimpleLogNormalDistribution() | ||
| 198 | : SimpleLogNormalDistribution(0.0) | ||
| 199 | {} | ||
| 200 | |||
| 201 | void param(const Parameters& p) | ||
| 202 | { normal_.param(p); } | ||
| 203 | |||
| 204 | Parameters param() const | ||
| 205 | { return normal_.param(); } | ||
| 206 | |||
| 207 | Scalar m() const { return normal_.m(); } | ||
| 208 | Scalar s() const { return normal_.s(); } | ||
| 209 | |||
| 210 | template<class Generator> | ||
| 211 | Scalar operator()(Generator& gen) | ||
| 212 | { | ||
| 213 | using std::exp; | ||
| 214 |
3/6✓ Branch 1 taken 15396 times.
✗ Branch 2 not taken.
✓ Branch 4 taken 69 times.
✗ Branch 5 not taken.
✓ Branch 7 taken 24 times.
✗ Branch 8 not taken.
|
17989 | return exp(normal_(gen)); |
| 215 | } | ||
| 216 | |||
| 217 | private: | ||
| 218 | SimpleNormalDistribution<Scalar> normal_; | ||
| 219 | }; | ||
| 220 | |||
| 221 | } // end namespace Dumux | ||
| 222 | |||
| 223 | #endif | ||
| 224 |