IT++ Logo
random.cpp
Go to the documentation of this file.
1 
29 #include <itpp/base/random.h>
30 #include <itpp/base/itcompat.h>
32 
33 
34 namespace itpp
35 {
36 
37 namespace random_details
38 {
39 
40 //Thread-local context for thread-safe RNGs
41 static ActiveDSFMT::Context thread_local_context;
42 #pragma omp threadprivate(thread_local_context)
43 //Thread-local context initialization flag
44 static bool is_thread_local_context_initialized = false;
45 #pragma omp threadprivate(is_thread_local_context_initialized)
46 
48 {
49  return thread_local_context;
50 }
51 
53 {
54  return is_thread_local_context_initialized;
55 }
56 
58 {
59  is_thread_local_context_initialized = true;
60 }
61 
68 static unsigned int hash_time_to_seed(time_t t, clock_t c)
69 {
70  static unsigned int differ = 0; // guarantee time-based seeds will change
71 
72  unsigned int h1 = 0;
73  unsigned char *p = (unsigned char *) &t;
74  for(size_t i = 0; i < sizeof(t); ++i) {
76  h1 += p[i];
77  }
78  unsigned int h2 = 0;
79  p = (unsigned char *) &c;
80  for(size_t j = 0; j < sizeof(c); ++j) {
82  h2 += p[j];
83  }
84  return (h1 + differ++) ^ h2;
85 }
86 
87 
88 
89 // ----------------------------------------------------------------------
90 // ActiveDSFMT (DSFMT_19937_RNG)
91 // ----------------------------------------------------------------------
92 template <>
93 const bool ActiveDSFMT::bigendian = is_bigendian();
94 
95 #if defined(__SSE2__)
96 template <>
97 const __m128i ActiveDSFMT::sse2_param_mask = _mm_set_epi32(ActiveDSFMT::MSK32_3, ActiveDSFMT::MSK32_4, ActiveDSFMT::MSK32_1, ActiveDSFMT::MSK32_2);
98 #endif // __SSE2__
99 }
100 
101 /*
102  *Global Seed Provider class definition.
103  *
104  * Provides unique seeds for thread-safe generators running in each thread.
105  */
106 class GlobalSeedProvider
107 {
108  static const unsigned int default_first_seed = 4257U;
109  typedef random_details::ActiveDSFMT DSFMT;
110 public:
111  //constructor
112  GlobalSeedProvider(): _dsfmt(_c), _first_seed_given(false) {
113  _dsfmt.init_gen_rand(default_first_seed);
114  }
115  //set new seed
116  void set_seed(unsigned int s) {_dsfmt.init_gen_rand(s);}
117  //get preset seed
118  unsigned int get_seed() {return _c.last_seed;}
119  //reset provider state to previously set one
120  void reset() { if(_first_seed_given) _dsfmt.init_gen_rand(get_seed());}
121  //set previously saved state from ivec
122  void set_state(const ivec& st) {
123  int size = (DSFMT::N + 1) * 4;
124  it_assert(st.size() == size + 1, "GlobalSeedProvider::state(): "
125  "Invalid state initialization vector");
126  uint32_t *psfmt = &_c.status[0].u32[0];
127  for(int i = 0; i < size; ++i) {
128  psfmt[i] = static_cast<uint32_t>(st(i));
129  }
130  _c.idx = st(size);
131  _first_seed_given = true;
132  }
133  //get current provider state
134  ivec get_state() {
135  int size = (DSFMT::N + 1) * 4;
136  uint32_t *psfmt = &_c.status[0].u32[0];
137  ivec state(size + 1); // size + 1 to save idx variable in the same vec
138  for(int i = 0; i < size; ++i) {
139  state(i) = static_cast<int>(psfmt[i]);
140  }
141  state(size) = _c.idx;
142  return state;
143  }
144  //randomize current provider state with system time
145  void randomize() {_dsfmt.init_gen_rand(random_details::hash_time_to_seed(time(0), clock())); _first_seed_given = true;}
146  //generate new seed for random number generators
147  unsigned int generate() {
148  if(_first_seed_given)
149  return _dsfmt.genrand_uint32();
150  else {
151  //return default seed on first request.
152  //it is done in order not to breake the old-style itpp tests.
153  //Some tests rely on the default state equal to 4257U
154  _first_seed_given = true;
155  return default_first_seed;
156  }
157  }
158 private:
159  //
160  DSFMT _dsfmt;
161  //
162  DSFMT::Context _c;
163  //
164  bool _first_seed_given;
165 };
166 
167 
168 //Global seed provider instance
169 GlobalSeedProvider& global_seed_provider()
170 {
171  static GlobalSeedProvider global_seed_provider_instance;
172  return global_seed_provider_instance;
173 }
174 
175 
176 void GlobalRNG_reset(unsigned int seed)
177 {
178  #pragma omp critical
179  {
180  global_seed_provider().set_seed(seed);
181  }
182 }
183 
184 
186 {
187  #pragma omp critical
188  {
189  global_seed_provider().reset();
190  }
191 }
192 
194 {
195  unsigned int s;
196  #pragma omp critical
197  {
198  s = global_seed_provider().generate();
199  }
200  return s;
201 }
202 
203 
205 {
206  #pragma omp critical
207  {
208  global_seed_provider().randomize();
209  }
210 }
211 
212 
213 void GlobalRNG_get_state(ivec &state)
214 {
215  #pragma omp critical
216  {
217  state = global_seed_provider().get_state();
218  }
219 }
220 
221 
222 void GlobalRNG_set_state(const ivec &state)
223 {
224  #pragma omp critical
225  {
226  global_seed_provider().set_state(state);
227  }
228 }
229 
230 void RNG_reset(unsigned int seed)
231 {
233  dsfmt.init_gen_rand(seed);
235 }
236 
237 
238 void RNG_reset()
239 {
241 
243  //already initialized. Reinit with last set seed;
244  dsfmt.init_gen_rand(random_details::lc_get().last_seed);
245  }
246  else {
247  //query global seed provider for new seed and init with it
250  }
251 }
252 
253 
255 {
257  dsfmt.init_gen_rand(random_details::hash_time_to_seed(time(0), clock()));
259 }
260 
261 
262 void RNG_get_state(ivec &state)
263 {
264  int size = (random_details::ActiveDSFMT::N + 1) * 4;
265  uint32_t *psfmt = &random_details::lc_get().status[0].u32[0];
266  state.set_size(size + 1); // size + 1 to save idx variable in the same vec
267  for(int i = 0; i < size; ++i) {
268  state(i) = static_cast<int>(psfmt[i]);
269  }
270  state(size) = random_details::lc_get().idx;
271 }
272 
273 
274 void RNG_set_state(const ivec &state)
275 {
276  int size = (random_details::ActiveDSFMT::N + 1) * 4;
277  it_assert(state.size() == size + 1, "RNG_set_state: "
278  "Invalid state initialization vector");
279  uint32_t *psfmt = &random_details::lc_get().status[0].u32[0];
280  for(int i = 0; i < size; ++i) {
281  psfmt[i] = static_cast<uint32_t>(state(i));
282  }
283  random_details::lc_get().idx = state(size);
284 }
285 
287 // I_Uniform_RNG
289 
291 {
292  setup(min, max);
293 }
294 
296 {
297  if(min <= max) {
298  lo = min;
299  hi = max;
300  }
301  else {
302  lo = max;
303  hi = min;
304  }
305 }
306 
307 void I_Uniform_RNG::get_setup(int &min, int &max) const
308 {
309  min = lo;
310  max = hi;
311 }
312 
314 {
315  ivec vv(n);
316 
317  for(int i = 0; i < n; i++)
318  vv(i) = sample();
319 
320  return vv;
321 }
322 
323 imat I_Uniform_RNG::operator()(int h, int w)
324 {
325  imat mm(h, w);
326  int i, j;
327 
328  for(i = 0; i < h; i++)
329  for(j = 0; j < w; j++)
330  mm(i, j) = sample();
331 
332  return mm;
333 }
334 
336 // Uniform_RNG
338 
340 {
341  setup(min, max);
342 }
343 
344 void Uniform_RNG::setup(double min, double max)
345 {
346  if(min <= max) {
347  lo_bound = min;
348  hi_bound = max;
349  }
350  else {
351  lo_bound = max;
352  hi_bound = min;
353  }
354 }
355 
356 void Uniform_RNG::get_setup(double &min, double &max) const
357 {
358  min = lo_bound;
359  max = hi_bound;
360 }
361 
363 // Exp_RNG
365 
367 {
368  setup(lambda);
369 }
370 
372 {
373  vec vv(n);
374 
375  for(int i = 0; i < n; i++)
376  vv(i) = sample();
377 
378  return vv;
379 }
380 
382 {
383  mat mm(h, w);
384  int i, j;
385 
386  for(i = 0; i < h; i++)
387  for(j = 0; j < w; j++)
388  mm(i, j) = sample();
389 
390  return mm;
391 }
392 
394 // Gamma_RNG
396 void Gamma_RNG::init_state()
397 {
398  const static double sqrt32 = 5.656854;
399 
400  const static double q1 = 0.04166669;
401  const static double q2 = 0.02083148;
402  const static double q3 = 0.00801191;
403  const static double q4 = 0.00144121;
404  const static double q5 = -7.388e-5;
405  const static double q6 = 2.4511e-4;
406  const static double q7 = 2.424e-4;
407 
408  double r = 1.0 / alpha;
409  _scale = 1.0 / beta;
410 
411  it_error_if(!std::isfinite(alpha) || !std::isfinite(_scale) || (alpha < 0.0)
412  || (_scale <= 0.0), "Gamma_RNG::init_state() - wrong parameters");
413 
414  _s2 = alpha - 0.5;
415  _s = std::sqrt(_s2);
416  _d = sqrt32 - _s * 12.0;
417 
418  _q0 = ((((((q7 * r + q6) * r + q5) * r + q4) * r + q3) * r
419  + q2) * r + q1) * r;
420 
421  /* Approximation depending on size of parameter alpha */
422  /* The constants in the expressions for _b, _si and _c */
423  /* were established by numerical experiments */
424  if(alpha <= 3.686) {
425  _b = 0.463 + _s + 0.178 * _s2;
426  _si = 1.235;
427  _c = 0.195 / _s - 0.079 + 0.16 * _s;
428  }
429  else if(alpha <= 13.022) {
430  _b = 1.654 + 0.0076 * _s2;
431  _si = 1.68 / _s + 0.275;
432  _c = 0.062 / _s + 0.024;
433  }
434  else {
435  _b = 1.77;
436  _si = 0.75;
437  _c = 0.1515 / _s;
438  }
439 
440 }
442 {
443  vec vv(n);
444  for(int i = 0; i < n; i++)
445  vv(i) = sample();
446  return vv;
447 }
448 
449 mat Gamma_RNG::operator()(int r, int c)
450 {
451  mat mm(r, c);
452  for(int i = 0; i < r * c; i++)
453  mm(i) = sample();
454  return mm;
455 }
456 
458 {
459  // A copy of rgamma code from the R package, adapted to IT++ by Vasek
460  // Smidl
461 
462  /* Constants : */
463  const static double exp_m1 = 0.36787944117144232159;/* exp(-1) = 1/e */
464 
465  const static double a1 = 0.3333333;
466  const static double a2 = -0.250003;
467  const static double a3 = 0.2000062;
468  const static double a4 = -0.1662921;
469  const static double a5 = 0.1423657;
470  const static double a6 = -0.1367177;
471  const static double a7 = 0.1233795;
472 
473  double e, p, q, t, u, v, w, x, ret_val;
474  double a = alpha;
475  double scale = _scale;
476 
477  if(a < 1.) { /* GS algorithm for parameters a < 1 */
478  if(a == 0)
479  return 0.;
480  e = 1.0 + exp_m1 * a;
481  for(;;) { //VS repeat
482  p = e * RNG.genrand_open_open();
483  if(p >= 1.0) {
484  x = -std::log((e - p) / a);
485  if(-std::log(RNG.genrand_open_close()) >= (1.0 - a) * std::log(x))
486  break;
487  }
488  else {
489  x = std::exp(std::log(p) / a);
490  if(-std::log(RNG.genrand_open_close()) >= x)
491  break;
492  }
493  }
494  return scale * x;
495  }
496 
497  /* --- a >= 1 : GD algorithm --- */
498 
499  /* Step 1: t = standard normal deviate, x = (s,1/2) -normal deviate. */
500  /* immediate acceptance (i) */
501  t = NRNG.sample();
502  x = _s + 0.5 * t;
503  ret_val = x * x;
504  if(t >= 0.0)
505  return scale * ret_val;
506 
507  /* Step 2: u = 0,1 - uniform sample. squeeze acceptance (s) */
508  u = RNG.genrand_close_open();
509  if((_d * u) <= (t * t * t))
510  return scale * ret_val;
511 
512 
513  /* Step 3: no quotient test if x not positive */
514  if(x > 0.0) {
515  /* Step 4: calculation of v and quotient q */
516  v = t / (_s + _s);
517  if(std::fabs(v) <= 0.25)
518  q = _q0 + 0.5 * t * t * ((((((a7 * v + a6) * v + a5) * v + a4) * v
519  + a3) * v + a2) * v + a1) * v;
520  else
521  q = _q0 - _s * t + 0.25 * t * t + (_s2 + _s2) * log(1.0 + v);
522 
523  /* Step 5: quotient acceptance (q) */
524  if(log(1.0 - u) <= q)
525  return scale * ret_val;
526  }
527 
528  for(;;) { //VS repeat
529  /* Step 6: e = standard exponential deviate
530  * u = 0,1 -uniform deviate
531  * t = (b,si)-double exponential (laplace) sample */
532  e = -std::log(RNG.genrand_open_close()); //see Exponential_RNG
533  u = RNG.genrand_open_close();
534  u = u + u - 1.0;
535  if(u < 0.0)
536  t = _b - _si * e;
537  else
538  t = _b + _si * e;
539  /* Step 7: rejection if t < tau(1) = -0.71874483771719 */
540  if(t >= -0.71874483771719) {
541  /* Step 8: calculation of v and quotient q */
542  v = t / (_s + _s);
543  if(std::fabs(v) <= 0.25)
544  q = _q0 + 0.5 * t * t *
545  ((((((a7 * v + a6) * v + a5) * v + a4) * v + a3) * v
546  + a2) * v + a1) * v;
547  else
548  q = _q0 - _s * t + 0.25 * t * t + (_s2 + _s2) * log(1.0 + v);
549  /* Step 9: hat acceptance (h) */
550  /* (if q not positive go to step 6) */
551  if(q > 0.0) {
552  // Try to use w = expm1(q); (Not supported on w32)
553  w = expm1(q);
554  /* ^^^^^ original code had approximation with rel.err < 2e-7 */
555  /* if t is rejected sample again at step 6 */
556  if((_c * std::fabs(u)) <= (w * std::exp(e - 0.5 * t * t)))
557  break;
558  }
559  }
560  } /* repeat .. until `t' is accepted */
561  x = _s + 0.5 * t;
562  return scale * x * x;
563 }
564 
566 // Normal_RNG
568 
569 void Normal_RNG::get_setup(double &meanval, double &variance) const
570 {
571  meanval = mean;
572  variance = sigma * sigma;
573 }
574 
575 // (Ziggurat) tabulated values for the heigt of the Ziggurat levels
576 const double Normal_RNG::ytab[128] = {
577  1, 0.963598623011, 0.936280813353, 0.913041104253,
578  0.892278506696, 0.873239356919, 0.855496407634, 0.838778928349,
579  0.822902083699, 0.807732738234, 0.793171045519, 0.779139726505,
580  0.765577436082, 0.752434456248, 0.739669787677, 0.727249120285,
581  0.715143377413, 0.703327646455, 0.691780377035, 0.68048276891,
582  0.669418297233, 0.65857233912, 0.647931876189, 0.637485254896,
583  0.62722199145, 0.617132611532, 0.607208517467, 0.597441877296,
584  0.587825531465, 0.578352913803, 0.569017984198, 0.559815170911,
585  0.550739320877, 0.541785656682, 0.532949739145, 0.524227434628,
586  0.515614886373, 0.507108489253, 0.498704867478, 0.490400854812,
587  0.482193476986, 0.47407993601, 0.466057596125, 0.458123971214,
588  0.450276713467, 0.442513603171, 0.434832539473, 0.427231532022,
589  0.419708693379, 0.41226223212, 0.404890446548, 0.397591718955,
590  0.390364510382, 0.383207355816, 0.376118859788, 0.369097692334,
591  0.362142585282, 0.355252328834, 0.348425768415, 0.341661801776,
592  0.334959376311, 0.328317486588, 0.321735172063, 0.31521151497,
593  0.308745638367, 0.302336704338, 0.29598391232, 0.289686497571,
594  0.283443729739, 0.27725491156, 0.271119377649, 0.265036493387,
595  0.259005653912, 0.253026283183, 0.247097833139, 0.241219782932,
596  0.235391638239, 0.229612930649, 0.223883217122, 0.218202079518,
597  0.212569124201, 0.206983981709, 0.201446306496, 0.195955776745,
598  0.190512094256, 0.185114984406, 0.179764196185, 0.174459502324,
599  0.169200699492, 0.1639876086, 0.158820075195, 0.153697969964,
600  0.148621189348, 0.143589656295, 0.138603321143, 0.133662162669,
601  0.128766189309, 0.123915440582, 0.119109988745, 0.114349940703,
602  0.10963544023, 0.104966670533, 0.100343857232, 0.0957672718266,
603  0.0912372357329, 0.0867541250127, 0.082318375932, 0.0779304915295,
604  0.0735910494266, 0.0693007111742, 0.065060233529, 0.0608704821745,
605  0.056732448584, 0.05264727098, 0.0486162607163, 0.0446409359769,
606  0.0407230655415, 0.0368647267386, 0.0330683839378, 0.0293369977411,
607  0.0256741818288, 0.0220844372634, 0.0185735200577, 0.0151490552854,
608  0.0118216532614, 0.00860719483079, 0.00553245272614, 0.00265435214565
609 };
610 
611 /*
612  * (Ziggurat) tabulated values for 2^24 times x[i]/x[i+1], used to accept
613  * for U*x[i+1]<=x[i] without any floating point operations
614  */
615 const unsigned int Normal_RNG::ktab[128] = {
616  0, 12590644, 14272653, 14988939,
617  15384584, 15635009, 15807561, 15933577,
618  16029594, 16105155, 16166147, 16216399,
619  16258508, 16294295, 16325078, 16351831,
620  16375291, 16396026, 16414479, 16431002,
621  16445880, 16459343, 16471578, 16482744,
622  16492970, 16502368, 16511031, 16519039,
623  16526459, 16533352, 16539769, 16545755,
624  16551348, 16556584, 16561493, 16566101,
625  16570433, 16574511, 16578353, 16581977,
626  16585398, 16588629, 16591685, 16594575,
627  16597311, 16599901, 16602354, 16604679,
628  16606881, 16608968, 16610945, 16612818,
629  16614592, 16616272, 16617861, 16619363,
630  16620782, 16622121, 16623383, 16624570,
631  16625685, 16626730, 16627708, 16628619,
632  16629465, 16630248, 16630969, 16631628,
633  16632228, 16632768, 16633248, 16633671,
634  16634034, 16634340, 16634586, 16634774,
635  16634903, 16634972, 16634980, 16634926,
636  16634810, 16634628, 16634381, 16634066,
637  16633680, 16633222, 16632688, 16632075,
638  16631380, 16630598, 16629726, 16628757,
639  16627686, 16626507, 16625212, 16623794,
640  16622243, 16620548, 16618698, 16616679,
641  16614476, 16612071, 16609444, 16606571,
642  16603425, 16599973, 16596178, 16591995,
643  16587369, 16582237, 16576520, 16570120,
644  16562917, 16554758, 16545450, 16534739,
645  16522287, 16507638, 16490152, 16468907,
646  16442518, 16408804, 16364095, 16301683,
647  16207738, 16047994, 15704248, 15472926
648 };
649 
650 // (Ziggurat) tabulated values of 2^{-24}*x[i]
651 const double Normal_RNG::wtab[128] = {
652  1.62318314817e-08, 2.16291505214e-08, 2.54246305087e-08, 2.84579525938e-08,
653  3.10340022482e-08, 3.33011726243e-08, 3.53439060345e-08, 3.72152672658e-08,
654  3.8950989572e-08, 4.05763964764e-08, 4.21101548915e-08, 4.35664624904e-08,
655  4.49563968336e-08, 4.62887864029e-08, 4.75707945735e-08, 4.88083237257e-08,
656  5.00063025384e-08, 5.11688950428e-08, 5.22996558616e-08, 5.34016475624e-08,
657  5.44775307871e-08, 5.55296344581e-08, 5.65600111659e-08, 5.75704813695e-08,
658  5.85626690412e-08, 5.95380306862e-08, 6.04978791776e-08, 6.14434034901e-08,
659  6.23756851626e-08, 6.32957121259e-08, 6.42043903937e-08, 6.51025540077e-08,
660  6.59909735447e-08, 6.68703634341e-08, 6.77413882848e-08, 6.8604668381e-08,
661  6.94607844804e-08, 7.03102820203e-08, 7.11536748229e-08, 7.1991448372e-08,
662  7.2824062723e-08, 7.36519550992e-08, 7.44755422158e-08, 7.52952223703e-08,
663  7.61113773308e-08, 7.69243740467e-08, 7.77345662086e-08, 7.85422956743e-08,
664  7.93478937793e-08, 8.01516825471e-08, 8.09539758128e-08, 8.17550802699e-08,
665  8.25552964535e-08, 8.33549196661e-08, 8.41542408569e-08, 8.49535474601e-08,
666  8.57531242006e-08, 8.65532538723e-08, 8.73542180955e-08, 8.8156298059e-08,
667  8.89597752521e-08, 8.97649321908e-08, 9.05720531451e-08, 9.138142487e-08,
668  9.21933373471e-08, 9.30080845407e-08, 9.38259651738e-08, 9.46472835298e-08,
669  9.54723502847e-08, 9.63014833769e-08, 9.71350089201e-08, 9.79732621669e-08,
670  9.88165885297e-08, 9.96653446693e-08, 1.00519899658e-07, 1.0138063623e-07,
671  1.02247952126e-07, 1.03122261554e-07, 1.04003996769e-07, 1.04893609795e-07,
672  1.05791574313e-07, 1.06698387725e-07, 1.07614573423e-07, 1.08540683296e-07,
673  1.09477300508e-07, 1.1042504257e-07, 1.11384564771e-07, 1.12356564007e-07,
674  1.13341783071e-07, 1.14341015475e-07, 1.15355110887e-07, 1.16384981291e-07,
675  1.17431607977e-07, 1.18496049514e-07, 1.19579450872e-07, 1.20683053909e-07,
676  1.21808209468e-07, 1.2295639141e-07, 1.24129212952e-07, 1.25328445797e-07,
677  1.26556042658e-07, 1.27814163916e-07, 1.29105209375e-07, 1.30431856341e-07,
678  1.31797105598e-07, 1.3320433736e-07, 1.34657379914e-07, 1.36160594606e-07,
679  1.37718982103e-07, 1.39338316679e-07, 1.41025317971e-07, 1.42787873535e-07,
680  1.44635331499e-07, 1.4657889173e-07, 1.48632138436e-07, 1.50811780719e-07,
681  1.53138707402e-07, 1.55639532047e-07, 1.58348931426e-07, 1.61313325908e-07,
682  1.64596952856e-07, 1.68292495203e-07, 1.72541128694e-07, 1.77574279496e-07,
683  1.83813550477e-07, 1.92166040885e-07, 2.05295471952e-07, 2.22600839893e-07
684 };
685 
686 // (Ziggurat) position of right-most step
687 const double Normal_RNG::PARAM_R = 3.44428647676;
688 
689 // Get a Normal distributed (0,1) sample
691 {
692  uint32_t u, sign, i, j;
693  double x, y;
694 
695  while(true) {
696  u = RNG.genrand_uint32();
697  sign = u & 0x80; // 1 bit for the sign
698  i = u & 0x7f; // 7 bits to choose the step
699  j = u >> 8; // 24 bits for the x-value
700 
701  x = j * wtab[i];
702 
703  if(j < ktab[i])
704  break;
705 
706  if(i < 127) {
707  y = ytab[i + 1] + (ytab[i] - ytab[i + 1]) * RNG.genrand_close_open();
708  }
709  else {
710  x = PARAM_R - std::log(1.0 - RNG.genrand_close_open()) / PARAM_R;
711  y = std::exp(-PARAM_R * (x - 0.5 * PARAM_R)) * RNG.genrand_close_open();
712  }
713 
714  if(y < std::exp(-0.5 * x * x))
715  break;
716  }
717  return sign ? x : -x;
718 }
719 
720 
722 // Laplace_RNG
724 
725 Laplace_RNG::Laplace_RNG(double meanval, double variance)
726 {
727  setup(meanval, variance);
728 }
729 
730 void Laplace_RNG::setup(double meanval, double variance)
731 {
732  mean = meanval;
733  var = variance;
734  sqrt_12var = std::sqrt(variance / 2.0);
735 }
736 
737 void Laplace_RNG::get_setup(double &meanval, double &variance) const
738 {
739  meanval = mean;
740  variance = var;
741 }
742 
743 
744 
746 {
747  vec vv(n);
748 
749  for(int i = 0; i < n; i++)
750  vv(i) = sample();
751 
752  return vv;
753 }
754 
755 mat Laplace_RNG::operator()(int h, int w)
756 {
757  mat mm(h, w);
758  int i, j;
759 
760  for(i = 0; i < h; i++)
761  for(j = 0; j < w; j++)
762  mm(i, j) = sample();
763 
764  return mm;
765 }
766 
768 // AR1_Normal_RNG
770 
771 AR1_Normal_RNG::AR1_Normal_RNG(double meanval, double variance, double rho)
772 {
773  setup(meanval, variance, rho);
774 }
775 
776 void AR1_Normal_RNG::setup(double meanval, double variance, double rho)
777 {
778  mean = meanval;
779  var = variance;
780  r = rho;
781  factr = -2.0 * var * (1.0 - rho * rho);
782  mem = 0.0;
783  odd = true;
784 }
785 
786 void AR1_Normal_RNG::get_setup(double &meanval, double &variance,
787  double &rho) const
788 {
789  meanval = mean;
790  variance = var;
791  rho = r;
792 }
793 
795 {
796  vec vv(n);
797 
798  for(int i = 0; i < n; i++)
799  vv(i) = sample();
800 
801  return vv;
802 }
803 
804 mat AR1_Normal_RNG::operator()(int h, int w)
805 {
806  mat mm(h, w);
807  int i, j;
808 
809  for(i = 0; i < h; i++)
810  for(j = 0; j < w; j++)
811  mm(i, j) = sample();
812 
813  return mm;
814 }
815 
817 {
818  mem = 0.0;
819 }
820 
822 // Weibull_RNG
824 
825 Weibull_RNG::Weibull_RNG(double lambda, double beta)
826 {
827  setup(lambda, beta);
828 }
829 
830 void Weibull_RNG::setup(double lambda, double beta)
831 {
832  l = lambda;
833  b = beta;
834  mean = tgamma(1.0 + 1.0 / b) / l;
835  var = tgamma(1.0 + 2.0 / b) / (l * l) - mean;
836 }
837 
838 
840 {
841  vec vv(n);
842 
843  for(int i = 0; i < n; i++)
844  vv(i) = sample();
845 
846  return vv;
847 }
848 
849 mat Weibull_RNG::operator()(int h, int w)
850 {
851  mat mm(h, w);
852  int i, j;
853 
854  for(i = 0; i < h; i++)
855  for(j = 0; j < w; j++)
856  mm(i, j) = sample();
857 
858  return mm;
859 }
860 
862 // Rayleigh_RNG
864 
866 {
867  setup(sigma);
868 }
869 
871 {
872  vec vv(n);
873 
874  for(int i = 0; i < n; i++)
875  vv(i) = sample();
876 
877  return vv;
878 }
879 
880 mat Rayleigh_RNG::operator()(int h, int w)
881 {
882  mat mm(h, w);
883  int i, j;
884 
885  for(i = 0; i < h; i++)
886  for(j = 0; j < w; j++)
887  mm(i, j) = sample();
888 
889  return mm;
890 }
891 
893 // Rice_RNG
895 
896 Rice_RNG::Rice_RNG(double lambda, double beta)
897 {
898  setup(lambda, beta);
899 }
900 
902 {
903  vec vv(n);
904 
905  for(int i = 0; i < n; i++)
906  vv(i) = sample();
907 
908  return vv;
909 }
910 
911 mat Rice_RNG::operator()(int h, int w)
912 {
913  mat mm(h, w);
914  int i, j;
915 
916  for(i = 0; i < h; i++)
917  for(j = 0; j < w; j++)
918  mm(i, j) = sample();
919 
920  return mm;
921 }
922 
923 } // namespace itpp
SourceForge Logo

Generated on Sat Jul 6 2013 10:54:21 for IT++ by Doxygen 1.8.2