1 | #ifndef _theplu_yat_random_ |
---|
2 | #define _theplu_yat_random_ |
---|
3 | |
---|
4 | // $Id: random.h 3518 2016-10-05 08:01:11Z peter $ |
---|
5 | |
---|
6 | /* |
---|
7 | Copyright (C) 2005, 2006, 2007, 2008 Jari Häkkinen, Peter Johansson |
---|
8 | Copyright (C) 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016 Peter Johansson |
---|
9 | |
---|
10 | This file is part of the yat library, http://dev.thep.lu.se/yat |
---|
11 | |
---|
12 | The yat library is free software; you can redistribute it and/or |
---|
13 | modify it under the terms of the GNU General Public License as |
---|
14 | published by the Free Software Foundation; either version 3 of the |
---|
15 | License, or (at your option) any later version. |
---|
16 | |
---|
17 | The yat library is distributed in the hope that it will be useful, |
---|
18 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
19 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
---|
20 | General Public License for more details. |
---|
21 | |
---|
22 | You should have received a copy of the GNU General Public License |
---|
23 | along with yat. If not, see <http://www.gnu.org/licenses/>. |
---|
24 | */ |
---|
25 | |
---|
26 | #include "yat/statistics/Histogram.h" |
---|
27 | #include "yat/utility/deprecate.h" |
---|
28 | |
---|
29 | // always include this before <boost/excpetion_ptr.hpp> indirectly below |
---|
30 | #include "yat/utility/boost_exception_ptr.h" |
---|
31 | |
---|
32 | #include <boost/concept_check.hpp> |
---|
33 | #include <boost/iterator/iterator_concepts.hpp> |
---|
34 | #include <boost/thread.hpp> |
---|
35 | #include <boost/thread/tss.hpp> |
---|
36 | |
---|
37 | #include <gsl/gsl_rng.h> |
---|
38 | #include <gsl/gsl_randist.h> |
---|
39 | |
---|
40 | #include <algorithm> |
---|
41 | #include <string> |
---|
42 | #include <vector> |
---|
43 | |
---|
44 | namespace theplu { |
---|
45 | namespace yat { |
---|
46 | namespace random { |
---|
47 | |
---|
48 | //forward declaration |
---|
49 | class RNG_state; |
---|
50 | |
---|
51 | /// |
---|
52 | /// @brief Random Number Generator |
---|
53 | /// |
---|
54 | /// The RNG class is wrapper to the GSL random number generator |
---|
55 | /// (rng). In yat 0.8 (or older) this class provided a single global |
---|
56 | /// instance of the rng, and made sure there was only one point of |
---|
57 | /// access to the generator. Since version 0.9 this class provides |
---|
58 | /// one rng per thread in order to avoid collisions in multi-thread |
---|
59 | /// applications. |
---|
60 | /// |
---|
61 | /// There are many different rng's available in GSL. RNG uses the |
---|
62 | /// default generator, unless the global variable \c gsl_rng_default |
---|
63 | /// has been modified (see <a |
---|
64 | /// href=\gsl_url/Random-number-environment-variables.html>GSL |
---|
65 | /// Manual</a>). Note, \c gsl_rng_default should be changed before |
---|
66 | /// RNG creates its generator and safest way to achieve this is to |
---|
67 | /// modify \c gsl_rng_default prior calling instance() the first |
---|
68 | /// time. |
---|
69 | /// |
---|
70 | /// There is information about how to change seeding and generators |
---|
71 | /// at run time without recompilation using environment variables in |
---|
72 | /// the <a |
---|
73 | /// href=\gsl_url/Random-number-environment-variables.html>GSL |
---|
74 | /// Manual</a>. RNG supports seeding at compile time if you don't |
---|
75 | /// want to bother about environment variables and GSL. |
---|
76 | /// |
---|
77 | /// The class provides one generator per thread. The first generator |
---|
78 | /// created is seeded with \c gsl_rng_default_seed and subsequent |
---|
79 | /// generators are seeded with \c gsl_rng_default_seed + 1, \c |
---|
80 | /// gsl_rng_default_seed + 2 etc, unless the seed has been modified |
---|
81 | /// with seed() or seed_from_devurandom(). |
---|
82 | /// |
---|
83 | /// @see Design Patterns (the singleton and adapter pattern). <a |
---|
84 | /// href=\gsl_url/Random-Number-Generation.html>GSL documentation</a>. |
---|
85 | /// |
---|
86 | class RNG |
---|
87 | { |
---|
88 | public: |
---|
89 | |
---|
90 | /// |
---|
91 | /// @brief Get an instance of the Random Number Generator. |
---|
92 | /// |
---|
93 | /// Get an instance of RNG. If a random number generator is not |
---|
94 | /// already created for current thread, the call will create a new |
---|
95 | /// generator of type \c gsl_rng_default. If it is the first |
---|
96 | /// generator created it will be seeded with \c |
---|
97 | /// gsl_rng_default_seed; otherwise created generator will be |
---|
98 | /// seeded with \c seed + \c n, where \c seed is the latest seed |
---|
99 | /// set (with seed() or seed_from_devurandom()) The seed may be |
---|
100 | /// changed with the seed or seed_from_devurandom member |
---|
101 | /// functions. |
---|
102 | /// |
---|
103 | /// @return A pointer to the random number generator. |
---|
104 | /// |
---|
105 | /// @see seed and seed_from_devurandom |
---|
106 | /// |
---|
107 | static RNG* instance(void); |
---|
108 | |
---|
109 | /// |
---|
110 | /// @brief Returns the largest number that the random number |
---|
111 | /// generator can return. |
---|
112 | /// |
---|
113 | unsigned long max(void) const; |
---|
114 | |
---|
115 | /// |
---|
116 | /// @brief Returns the smallest number that the random number |
---|
117 | /// generator can return. |
---|
118 | /// |
---|
119 | unsigned long min(void) const; |
---|
120 | |
---|
121 | /// |
---|
122 | /// @brief Returns the name of the random number generator |
---|
123 | /// |
---|
124 | std::string name(void) const; |
---|
125 | |
---|
126 | /// |
---|
127 | /// Access underlying GSL random number generator speicific to |
---|
128 | /// current thread. Behaviour of returned generator is undefined |
---|
129 | /// outside current thread. |
---|
130 | /// |
---|
131 | /// @return const pointer to underlying GSL random generator. |
---|
132 | /// |
---|
133 | const gsl_rng* rng(void) const; |
---|
134 | |
---|
135 | /// |
---|
136 | /// @brief Set the seed \a s for the rng. |
---|
137 | /// |
---|
138 | /// Set the seed \a s for the rng. If \a s is zero, a default |
---|
139 | /// value from the rng's original implementation is used (cf. GSL |
---|
140 | /// documentation). |
---|
141 | /// |
---|
142 | /// This function will also effect generators created subsequently |
---|
143 | /// in other threads. The seed \a s is saved and subsequent |
---|
144 | /// generators will be created with seed \c s + 1, \c s + 2, etc. |
---|
145 | /// |
---|
146 | /// @see seed_from_devurandom |
---|
147 | /// |
---|
148 | void seed(unsigned long s) const; |
---|
149 | |
---|
150 | /// |
---|
151 | /// @brief Seed the rng using the /dev/urandom device. |
---|
152 | /// |
---|
153 | /// This function will also effect generators in other threads |
---|
154 | /// created subsequntly (see seed()). |
---|
155 | /// |
---|
156 | /// @return The seed acquired from /dev/urandom. |
---|
157 | /// |
---|
158 | unsigned long seed_from_devurandom(void); |
---|
159 | |
---|
160 | /** |
---|
161 | \brief Set the state to \a state. |
---|
162 | |
---|
163 | \return 0 always. |
---|
164 | |
---|
165 | \note this function only effects the RNG in current thread |
---|
166 | |
---|
167 | \throw utility::GSL_error on error |
---|
168 | |
---|
169 | \see gsl_rng_memcpy |
---|
170 | */ |
---|
171 | // return int for backward compatibility with yat 0.8 |
---|
172 | int set_state(const RNG_state&); |
---|
173 | |
---|
174 | private: |
---|
175 | RNG(void); |
---|
176 | |
---|
177 | /** |
---|
178 | \brief Not implemented. |
---|
179 | |
---|
180 | This copy contructor is not implemented. The constructor is |
---|
181 | declared in order to avoid compiler generated default copy |
---|
182 | constructor. |
---|
183 | */ |
---|
184 | RNG(const RNG&); |
---|
185 | |
---|
186 | /** |
---|
187 | There can be only one RNG so assignment is always |
---|
188 | self-assignment and we do not allow it |
---|
189 | */ |
---|
190 | RNG& operator=(const RNG&); |
---|
191 | |
---|
192 | virtual ~RNG(void); |
---|
193 | void rng_alloc(void) const; |
---|
194 | |
---|
195 | static RNG* instance_; |
---|
196 | // holds one gsl_rng per thread. Access through rng(void) so a |
---|
197 | // gsl_rng is allocated if necessary. |
---|
198 | mutable boost::thread_specific_ptr<gsl_rng> rng_; |
---|
199 | mutable unsigned long seed_; |
---|
200 | // guard needs to be mutable because major mission for it is to protect seed_ against multi-access, and seed_ is mutable... |
---|
201 | mutable boost::mutex mutex_; |
---|
202 | }; |
---|
203 | |
---|
204 | |
---|
205 | /// |
---|
206 | /// @brief Class holding state of a random generator |
---|
207 | /// |
---|
208 | class RNG_state |
---|
209 | { |
---|
210 | public: |
---|
211 | /// |
---|
212 | /// @brief Constructor |
---|
213 | /// |
---|
214 | explicit RNG_state(const RNG*); |
---|
215 | |
---|
216 | /** |
---|
217 | Copy Constructor |
---|
218 | |
---|
219 | \since Explicitely declared since yat 0.5 |
---|
220 | */ |
---|
221 | RNG_state(const RNG_state&); |
---|
222 | |
---|
223 | /// |
---|
224 | /// @brief Destructor |
---|
225 | /// |
---|
226 | ~RNG_state(void); |
---|
227 | |
---|
228 | /// |
---|
229 | /// @return const pointer to underlying GSL random generator. |
---|
230 | /// |
---|
231 | const gsl_rng* rng(void) const; |
---|
232 | |
---|
233 | /** |
---|
234 | Assignment operator |
---|
235 | |
---|
236 | \since Explicitely declared since yat 0.5 |
---|
237 | */ |
---|
238 | RNG_state& operator=(const RNG_state&); |
---|
239 | |
---|
240 | private: |
---|
241 | gsl_rng* rng_; |
---|
242 | |
---|
243 | void clone(const gsl_rng&); |
---|
244 | }; |
---|
245 | |
---|
246 | |
---|
247 | // --------------------- Discrete distribtuions --------------------- |
---|
248 | |
---|
249 | /// |
---|
250 | /// @brief Discrete random number distributions. |
---|
251 | /// |
---|
252 | /// Abstract base class for discrete random number |
---|
253 | /// distributions. Given K discrete events with different |
---|
254 | /// probabilities \f$ P[k] \f$, produce a random value k consistent |
---|
255 | /// with its probability. |
---|
256 | /// |
---|
257 | class Discrete |
---|
258 | { |
---|
259 | public: |
---|
260 | /** |
---|
261 | type returned by operator() |
---|
262 | |
---|
263 | \since New in yat 0.10 |
---|
264 | */ |
---|
265 | typedef unsigned long int result_type; |
---|
266 | |
---|
267 | /// |
---|
268 | /// @brief Constructor |
---|
269 | /// |
---|
270 | Discrete(void); |
---|
271 | |
---|
272 | /// |
---|
273 | /// @brief The destructor |
---|
274 | /// |
---|
275 | virtual ~Discrete(void); |
---|
276 | |
---|
277 | /// |
---|
278 | /// @brief Set the seed to \a s. |
---|
279 | /// |
---|
280 | /// Set the seed to \a s in the underlying rng. If \a s is zero, a |
---|
281 | /// default value from the rng's original implementation is used |
---|
282 | /// (cf. GSL documentation). |
---|
283 | /// |
---|
284 | /// \deprecated Provided for backward compatibility with the 0.7 |
---|
285 | /// API. Use RNG::instance()->seed(s) instead. |
---|
286 | /// |
---|
287 | void seed(unsigned long s) const YAT_DEPRECATE; |
---|
288 | |
---|
289 | /// |
---|
290 | /// @brief Set the seed using the /dev/urandom device. |
---|
291 | /// |
---|
292 | /// @return The seed acquired from /dev/urandom. |
---|
293 | /// |
---|
294 | /// \deprecated Provided for backward compatibility with the 0.7 |
---|
295 | /// API. Use RNG::instance()->seed_from_devurandom() instead. |
---|
296 | /// |
---|
297 | unsigned long seed_from_devurandom(void) YAT_DEPRECATE; |
---|
298 | |
---|
299 | /// |
---|
300 | /// @return A random number. |
---|
301 | /// |
---|
302 | virtual result_type operator()(void) const = 0; |
---|
303 | |
---|
304 | protected: |
---|
305 | /// GSL random gererator |
---|
306 | RNG* rng_; |
---|
307 | }; |
---|
308 | |
---|
309 | /** |
---|
310 | \brief Binomial distribution |
---|
311 | |
---|
312 | \see gsl_ran_binomial |
---|
313 | |
---|
314 | \since New in yat 0.10 |
---|
315 | */ |
---|
316 | class Binomial : public Discrete |
---|
317 | { |
---|
318 | public: |
---|
319 | /** |
---|
320 | \brief Constructor |
---|
321 | |
---|
322 | Create an object that generates random numbers from binomial |
---|
323 | distribution, the number of successes in \n trials each with |
---|
324 | success probability \a p. |
---|
325 | */ |
---|
326 | Binomial(double p, unsigned int n); |
---|
327 | |
---|
328 | /** |
---|
329 | \return number from binomial distrubtion as parametrized in constructor. |
---|
330 | */ |
---|
331 | unsigned long operator()(void) const; |
---|
332 | private: |
---|
333 | double p_; |
---|
334 | unsigned int n_; |
---|
335 | }; |
---|
336 | |
---|
337 | /// |
---|
338 | /// @brief General |
---|
339 | /// |
---|
340 | class DiscreteGeneral : public Discrete |
---|
341 | { |
---|
342 | public: |
---|
343 | /// |
---|
344 | /// @brief Constructor |
---|
345 | /// |
---|
346 | /// @param hist is a Histogram defining the probability distribution |
---|
347 | /// |
---|
348 | explicit DiscreteGeneral(const statistics::Histogram& hist); |
---|
349 | |
---|
350 | /** |
---|
351 | \brief Copy constructor |
---|
352 | |
---|
353 | \since Explicitely implemented in yat 0.5 |
---|
354 | */ |
---|
355 | DiscreteGeneral(const DiscreteGeneral&); |
---|
356 | |
---|
357 | /// |
---|
358 | /// @brief Destructor |
---|
359 | /// |
---|
360 | ~DiscreteGeneral(void); |
---|
361 | |
---|
362 | /// |
---|
363 | /// The generated number is an integer and proportional to the |
---|
364 | /// frequency in the corresponding histogram bin. In other words, |
---|
365 | /// the probability that 0 is returned is proportinal to the size |
---|
366 | /// of the first bin. |
---|
367 | /// |
---|
368 | /// @return A random number. |
---|
369 | /// |
---|
370 | unsigned long operator()(void) const; |
---|
371 | |
---|
372 | /** |
---|
373 | \brief Assignment operator |
---|
374 | |
---|
375 | \since Explicitely implemented in yat 0.5 |
---|
376 | */ |
---|
377 | DiscreteGeneral& operator=(const DiscreteGeneral&); |
---|
378 | |
---|
379 | private: |
---|
380 | void free(void); |
---|
381 | void preproc(void); |
---|
382 | |
---|
383 | gsl_ran_discrete_t* gen_; |
---|
384 | std::vector<double> p_; |
---|
385 | }; |
---|
386 | |
---|
387 | /** |
---|
388 | @brief Discrete uniform distribution |
---|
389 | |
---|
390 | Discrete uniform distribution also known as the "equally likely |
---|
391 | outcomes" distribution. Each outcome, in this case an integer |
---|
392 | from [0,n-1] , have equal probability to occur. |
---|
393 | |
---|
394 | Distribution function \f$ p(k) = \frac{1}{n+1} \f$ for \f$ 0 \le |
---|
395 | k < n \f$ \n |
---|
396 | Expectation value: \f$ \frac{n-1}{2} \f$ \n |
---|
397 | Variance: \f$ \frac{1}{12}(n-1)(n+1) \f$ |
---|
398 | */ |
---|
399 | class DiscreteUniform |
---|
400 | : public Discrete, |
---|
401 | public std::unary_function<unsigned long, unsigned long> |
---|
402 | { |
---|
403 | public: |
---|
404 | /** |
---|
405 | \brief Constructor. |
---|
406 | |
---|
407 | The generator will generate integers within the range \f$ |
---|
408 | [0,n-1] \f$. If \a n is zero, then the whole range of the |
---|
409 | underlying RNG will be used \f$ [min,max] \f$. Setting \a n to |
---|
410 | zero is the preferred way to sample the whole range of the |
---|
411 | underlying RNG, i.e. not setting \n to RNG.max. |
---|
412 | |
---|
413 | \throw If \a n is larger than the maximum number the underlying |
---|
414 | random number generator can return, then a GSL_error exception |
---|
415 | is thrown. |
---|
416 | */ |
---|
417 | explicit DiscreteUniform(unsigned long n=0); |
---|
418 | |
---|
419 | /** |
---|
420 | \brief Get a random number |
---|
421 | |
---|
422 | The returned integer is either in the range [RNG.min,RNG.max] |
---|
423 | or [0,n-1] depending on how the random number generator was |
---|
424 | created. |
---|
425 | |
---|
426 | \see DiscreteUniform(const unsigned long n=0) |
---|
427 | */ |
---|
428 | unsigned long operator()(void) const; |
---|
429 | |
---|
430 | /** |
---|
431 | \brief Get a random integer in the range \f$ [0,n-1] \f$. |
---|
432 | |
---|
433 | All integers in the range [0,n-1] are equally likely. This |
---|
434 | function should be avoided for sampling the whole range of the |
---|
435 | underlying RNG. |
---|
436 | |
---|
437 | \throw GSL_error if \a n is larger than the range of the |
---|
438 | underlying generator. |
---|
439 | */ |
---|
440 | unsigned long operator()(unsigned long n) const; |
---|
441 | |
---|
442 | private: |
---|
443 | unsigned long range_; |
---|
444 | }; |
---|
445 | |
---|
446 | |
---|
447 | /** |
---|
448 | @brief Geomtric Distribution |
---|
449 | |
---|
450 | The number of independent trials with probability \em p until the |
---|
451 | first success. |
---|
452 | |
---|
453 | Probability function \f$ p(k) = p (1-p)^(k-1) \f$ |
---|
454 | |
---|
455 | \since New in yat 0.14 |
---|
456 | */ |
---|
457 | class Geometric : public Discrete |
---|
458 | { |
---|
459 | public: |
---|
460 | /** |
---|
461 | \brief Constructor |
---|
462 | |
---|
463 | \param p is probability for success in one trial |
---|
464 | */ |
---|
465 | Geometric(double p); |
---|
466 | |
---|
467 | /* |
---|
468 | \return a number from Geomtric distribution |
---|
469 | */ |
---|
470 | unsigned long operator()(void) const; |
---|
471 | |
---|
472 | /** |
---|
473 | \return a number from Geomtric distribution with success rate \a p |
---|
474 | |
---|
475 | \note this operator ignores parameters set in Constructor |
---|
476 | */ |
---|
477 | unsigned long operator()(double p) const; |
---|
478 | |
---|
479 | private: |
---|
480 | double p_; |
---|
481 | }; |
---|
482 | |
---|
483 | |
---|
484 | /** |
---|
485 | If we have \a n1 samples of type 1 and \a n2 samples of type 2 |
---|
486 | and draw \a t samples with replacement, number of drawn samples |
---|
487 | of type 1 will follow the hyper geometric distribution. |
---|
488 | |
---|
489 | \since New in yat 0.14 |
---|
490 | */ |
---|
491 | class HyperGeometric : public Discrete |
---|
492 | { |
---|
493 | public: |
---|
494 | /** |
---|
495 | \brief Defaul constructor |
---|
496 | */ |
---|
497 | HyperGeometric(void); |
---|
498 | |
---|
499 | /** |
---|
500 | \brief Constructor |
---|
501 | \param n1 number of samples of type 1 |
---|
502 | \param n2 number of samples of type 2 |
---|
503 | \param t number of samples to draw |
---|
504 | */ |
---|
505 | HyperGeometric(unsigned int n1, unsigned int n2, unsigned int t); |
---|
506 | |
---|
507 | /** |
---|
508 | \return random number from hypergeometric distribution using |
---|
509 | parameters set in constructor. |
---|
510 | */ |
---|
511 | unsigned long int operator()(void) const; |
---|
512 | |
---|
513 | /** |
---|
514 | \return random number from hypergeometric distribution using |
---|
515 | parameters passed. |
---|
516 | */ |
---|
517 | unsigned long int operator()(unsigned int n1, unsigned int n2, |
---|
518 | unsigned int t) const; |
---|
519 | private: |
---|
520 | unsigned int n1_; |
---|
521 | unsigned int n2_; |
---|
522 | unsigned int t_; |
---|
523 | }; |
---|
524 | |
---|
525 | /** |
---|
526 | We have \a n1 samples of type 1 and \a n2 samples of type 2. |
---|
527 | Samples are drawn with replacement until \a t samles of type 2 |
---|
528 | are drawn. Then \a k, number of drawn samples of type 1, follows |
---|
529 | negative hypergeometric distribution. |
---|
530 | |
---|
531 | \since New in yat 0.14 |
---|
532 | |
---|
533 | \see HyperGeometric |
---|
534 | */ |
---|
535 | class NegativeHyperGeometric : public Discrete |
---|
536 | { |
---|
537 | public: |
---|
538 | /** |
---|
539 | \brief Default constructor |
---|
540 | */ |
---|
541 | NegativeHyperGeometric(void); |
---|
542 | |
---|
543 | /** |
---|
544 | \brief Constructor |
---|
545 | \param n1 number of samples of type 1 |
---|
546 | \param n2 number of samples of type 2 |
---|
547 | \param t number of samples of type 2 to draw |
---|
548 | */ |
---|
549 | NegativeHyperGeometric(unsigned int n1, unsigned int n2, unsigned int t); |
---|
550 | |
---|
551 | /** |
---|
552 | \return random number from negative hypergeometric distribution |
---|
553 | using parameters set in constructor. |
---|
554 | */ |
---|
555 | unsigned long int operator()(void) const; |
---|
556 | |
---|
557 | /** |
---|
558 | \return random number from negative hypergeometric distribution |
---|
559 | using parameters passed. |
---|
560 | */ |
---|
561 | unsigned long int operator()(unsigned int n1, unsigned int n2, |
---|
562 | unsigned int t) const; |
---|
563 | private: |
---|
564 | unsigned int n1_; |
---|
565 | unsigned int n2_; |
---|
566 | unsigned int t_; |
---|
567 | }; |
---|
568 | |
---|
569 | |
---|
570 | /** |
---|
571 | @brief Poisson Distribution |
---|
572 | |
---|
573 | Having a Poisson process (i.e. no memory), number of occurences |
---|
574 | within a given time window is Poisson distributed. This |
---|
575 | distribution is the limit of a Binomial distribution when number |
---|
576 | of attempts is large, and the probability for one attempt to be |
---|
577 | succesful is small (in such a way that the expected number of |
---|
578 | succesful attempts is \f$ m \f$. |
---|
579 | |
---|
580 | Probability function \f$ p(k) = e^{-m}\frac{m^k}{k!} \f$ for \f$ 0 \le |
---|
581 | k \f$ \n |
---|
582 | Expectation value: \f$ m \f$ \n |
---|
583 | Variance: \f$ m \f$ |
---|
584 | */ |
---|
585 | class Poisson : public Discrete |
---|
586 | { |
---|
587 | public: |
---|
588 | /// |
---|
589 | /// @brief Constructor |
---|
590 | /// |
---|
591 | /// @param m is expectation value |
---|
592 | /// |
---|
593 | explicit Poisson(const double m=1); |
---|
594 | |
---|
595 | /// |
---|
596 | /// @return A Poisson distributed number. |
---|
597 | /// |
---|
598 | unsigned long operator()(void) const; |
---|
599 | |
---|
600 | /// |
---|
601 | /// @return A Poisson distributed number with expectation value |
---|
602 | /// \a m |
---|
603 | /// |
---|
604 | /// @note this operator ignores parameters set in Constructor |
---|
605 | /// |
---|
606 | unsigned long operator()(const double m) const; |
---|
607 | |
---|
608 | private: |
---|
609 | double m_; |
---|
610 | }; |
---|
611 | |
---|
612 | // --------------------- Continuous distribtuions --------------------- |
---|
613 | |
---|
614 | /// |
---|
615 | /// @brief Continuous random number distributions. |
---|
616 | /// |
---|
617 | /// Abstract base class for continuous random number distributions. |
---|
618 | /// |
---|
619 | class Continuous |
---|
620 | { |
---|
621 | public: |
---|
622 | /** |
---|
623 | type returned by operator() |
---|
624 | |
---|
625 | \since New in yat 0.10 |
---|
626 | */ |
---|
627 | typedef double result_type; |
---|
628 | |
---|
629 | /// |
---|
630 | /// @brief Constructor |
---|
631 | /// |
---|
632 | Continuous(void); |
---|
633 | |
---|
634 | /// |
---|
635 | /// @brief The destructor |
---|
636 | /// |
---|
637 | virtual ~Continuous(void); |
---|
638 | |
---|
639 | /// |
---|
640 | /// @brief Set the seed to \a s. |
---|
641 | /// |
---|
642 | /// Set the seed to \a s in the underlying rng. If \a s is zero, a |
---|
643 | /// default value from the rng's original implementation is used |
---|
644 | /// (cf. GSL documentation). |
---|
645 | /// |
---|
646 | /// \deprecated Provided for backward compatibility with the 0.7 |
---|
647 | /// API. Use RNG::instance()->seed(s) instead. |
---|
648 | /// |
---|
649 | void seed(unsigned long s) const YAT_DEPRECATE; |
---|
650 | |
---|
651 | /// |
---|
652 | /// @brief Set the seed using the /dev/urandom device. |
---|
653 | /// |
---|
654 | /// @return The seed acquired from /dev/urandom. |
---|
655 | /// |
---|
656 | /// \deprecated Provided for backward compatibility with the 0.7 |
---|
657 | /// API. Use RNG::instance()->seed_from_devurandom() instead. |
---|
658 | /// |
---|
659 | unsigned long seed_from_devurandom(void) YAT_DEPRECATE; |
---|
660 | |
---|
661 | /// |
---|
662 | /// @return A random number |
---|
663 | /// |
---|
664 | virtual result_type operator()(void) const = 0; |
---|
665 | |
---|
666 | protected: |
---|
667 | /// pointer to GSL random generator |
---|
668 | RNG* rng_; |
---|
669 | }; |
---|
670 | |
---|
671 | // ContinuousUniform is declared before ContinuousGeneral to avoid |
---|
672 | // forward declaration |
---|
673 | /// |
---|
674 | /// @brief Uniform distribution |
---|
675 | /// |
---|
676 | /// Class for generating a random number from a uniform distribution |
---|
677 | /// in the range [0,1), i.e. zero is included but not 1. |
---|
678 | /// |
---|
679 | /// Distribution function \f$ f(x) = 1 \f$ for \f$ 0 \le x < 1 \f$ \n |
---|
680 | /// Expectation value: 0.5 \n |
---|
681 | /// Variance: \f$ \frac{1}{12} \f$ |
---|
682 | /// |
---|
683 | class ContinuousUniform : public Continuous |
---|
684 | { |
---|
685 | public: |
---|
686 | double operator()(void) const; |
---|
687 | }; |
---|
688 | |
---|
689 | /// |
---|
690 | /// @brief Generates numbers from a histogram in a continuous manner. |
---|
691 | /// |
---|
692 | class ContinuousGeneral : public Continuous |
---|
693 | { |
---|
694 | public: |
---|
695 | /// |
---|
696 | /// @brief Constructor |
---|
697 | /// |
---|
698 | /// @param hist is a Histogram defining the probability distribution |
---|
699 | /// |
---|
700 | explicit ContinuousGeneral(const statistics::Histogram& hist); |
---|
701 | |
---|
702 | /// |
---|
703 | /// The number is generated in a two step process. First the bin |
---|
704 | /// in the histogram is randomly selected (see |
---|
705 | /// DiscreteGeneral). Then a number is generated uniformly from |
---|
706 | /// the interval defined by the bin. |
---|
707 | /// |
---|
708 | /// @return A random number. |
---|
709 | /// |
---|
710 | double operator()(void) const; |
---|
711 | |
---|
712 | private: |
---|
713 | const DiscreteGeneral discrete_; |
---|
714 | const statistics::Histogram hist_; |
---|
715 | ContinuousUniform u_; |
---|
716 | }; |
---|
717 | |
---|
718 | /** |
---|
719 | \brief Generator of random numbers from an exponential |
---|
720 | distribution. |
---|
721 | |
---|
722 | The distribution function is \f$ f(x) = \frac{1}{m}\exp(-x/a) |
---|
723 | \f$ for \f$ x \f$ with the expectation value \f$ m \f$ and |
---|
724 | variance \f$ m^2 \f$ |
---|
725 | */ |
---|
726 | class Exponential : public Continuous |
---|
727 | { |
---|
728 | public: |
---|
729 | /// |
---|
730 | /// @brief Constructor |
---|
731 | /// |
---|
732 | /// @param m is the expectation value of the distribution. |
---|
733 | /// |
---|
734 | explicit Exponential(const double m=1); |
---|
735 | |
---|
736 | /// |
---|
737 | /// @return A random number from exponential distribution. |
---|
738 | /// |
---|
739 | double operator()(void) const; |
---|
740 | |
---|
741 | /// |
---|
742 | /// @return A random number from exponential distribution, with |
---|
743 | /// expectation value \a m |
---|
744 | /// |
---|
745 | /// @note This operator ignores parameters given in constructor. |
---|
746 | /// |
---|
747 | double operator()(const double m) const; |
---|
748 | |
---|
749 | private: |
---|
750 | double m_; |
---|
751 | }; |
---|
752 | |
---|
753 | /** |
---|
754 | @brief Gaussian distribution |
---|
755 | |
---|
756 | Class for generating a random number from a Gaussian distribution |
---|
757 | between zero and unity. Utilizes the Box-Muller algorithm, which |
---|
758 | needs two calls to random generator. |
---|
759 | |
---|
760 | Distribution function \f$ f(x) = |
---|
761 | \frac{1}{\sqrt{2\pi\sigma^2}}\exp(-\frac{(x-\mu)^2}{2\sigma^2}) |
---|
762 | \f$ \n |
---|
763 | Expectation value: \f$ \mu \f$ \n |
---|
764 | Variance: \f$ \sigma^2 \f$ |
---|
765 | */ |
---|
766 | class Gaussian : public Continuous |
---|
767 | { |
---|
768 | public: |
---|
769 | /// |
---|
770 | /// @brief Constructor |
---|
771 | /// |
---|
772 | /// @param s is the standard deviation \f$ \sigma \f$ of distribution |
---|
773 | /// @param m is the expectation value \f$ \mu \f$ of the distribution |
---|
774 | /// |
---|
775 | explicit Gaussian(const double s=1, const double m=0); |
---|
776 | |
---|
777 | /// |
---|
778 | /// @return A random Gaussian number |
---|
779 | /// |
---|
780 | double operator()(void) const; |
---|
781 | |
---|
782 | /// |
---|
783 | /// @return A random Gaussian number with standard deviation \a s |
---|
784 | /// and expectation value 0. |
---|
785 | /// |
---|
786 | /// @note this operator ignores parameters given in Constructor |
---|
787 | /// |
---|
788 | double operator()(const double s) const; |
---|
789 | |
---|
790 | /// |
---|
791 | /// @return A random Gaussian number with standard deviation \a s |
---|
792 | /// and expectation value \a m. |
---|
793 | /// |
---|
794 | /// @note this operator ignores parameters given in Constructor |
---|
795 | /// |
---|
796 | double operator()(const double s, const double m) const; |
---|
797 | |
---|
798 | private: |
---|
799 | double m_; |
---|
800 | double s_; |
---|
801 | }; |
---|
802 | |
---|
803 | /** |
---|
804 | \brief Convenience function to shuffle a range with singleton RNG. |
---|
805 | |
---|
806 | Wrapper around std::random_shuffle using DiscreteUniform as |
---|
807 | random generator and thereby using the underlying RNG class, |
---|
808 | which is singleton. |
---|
809 | |
---|
810 | Type Requirements: |
---|
811 | - RandomAccessIterator is \random_access_iterator |
---|
812 | */ |
---|
813 | template<typename RandomAccessIterator> |
---|
814 | void random_shuffle(RandomAccessIterator first, RandomAccessIterator last) |
---|
815 | { |
---|
816 | DiscreteUniform rnd; |
---|
817 | std::random_shuffle(first, last, rnd); |
---|
818 | } |
---|
819 | |
---|
820 | }}} // of namespace random, yat, and theplu |
---|
821 | |
---|
822 | #endif |
---|