source: src/RandomNumbers/RandomNumberDistribution_Encapsulation.hpp@ 3f9eba

Action_Thermostats Add_AtomRandomPerturbation Add_FitFragmentPartialChargesAction Add_RotateAroundBondAction Add_SelectAtomByNameAction Added_ParseSaveFragmentResults AddingActions_SaveParseParticleParameters Adding_Graph_to_ChangeBondActions Adding_MD_integration_tests Adding_ParticleName_to_Atom Adding_StructOpt_integration_tests AtomFragments Automaking_mpqc_open AutomationFragmentation_failures Candidate_v1.5.4 Candidate_v1.6.0 Candidate_v1.6.1 ChangeBugEmailaddress ChangingTestPorts ChemicalSpaceEvaluator CombiningParticlePotentialParsing Combining_Subpackages Debian_Package_split Debian_package_split_molecuildergui_only Disabling_MemDebug Docu_Python_wait EmpiricalPotential_contain_HomologyGraph EmpiricalPotential_contain_HomologyGraph_documentation Enable_parallel_make_install Enhance_userguide Enhanced_StructuralOptimization Enhanced_StructuralOptimization_continued Example_ManyWaysToTranslateAtom Exclude_Hydrogens_annealWithBondGraph FitPartialCharges_GlobalError Fix_BoundInBox_CenterInBox_MoleculeActions Fix_ChargeSampling_PBC Fix_ChronosMutex Fix_FitPartialCharges Fix_FitPotential_needs_atomicnumbers Fix_ForceAnnealing Fix_IndependentFragmentGrids Fix_ParseParticles Fix_ParseParticles_split_forward_backward_Actions Fix_PopActions Fix_QtFragmentList_sorted_selection Fix_Restrictedkeyset_FragmentMolecule Fix_StatusMsg Fix_StepWorldTime_single_argument Fix_Verbose_Codepatterns Fix_fitting_potentials Fixes ForceAnnealing_goodresults ForceAnnealing_oldresults ForceAnnealing_tocheck ForceAnnealing_with_BondGraph ForceAnnealing_with_BondGraph_continued ForceAnnealing_with_BondGraph_continued_betteresults ForceAnnealing_with_BondGraph_contraction-expansion FragmentAction_writes_AtomFragments FragmentMolecule_checks_bonddegrees GeometryObjects Gui_Fixes Gui_displays_atomic_force_velocity ImplicitCharges IndependentFragmentGrids IndependentFragmentGrids_IndividualZeroInstances IndependentFragmentGrids_IntegrationTest IndependentFragmentGrids_Sole_NN_Calculation JobMarket_RobustOnKillsSegFaults JobMarket_StableWorkerPool JobMarket_unresolvable_hostname_fix MoreRobust_FragmentAutomation ODR_violation_mpqc_open PartialCharges_OrthogonalSummation PdbParser_setsAtomName PythonUI_with_named_parameters QtGui_reactivate_TimeChanged_changes Recreated_GuiChecks Rewrite_FitPartialCharges RotateToPrincipalAxisSystem_UndoRedo SaturateAtoms_findBestMatching SaturateAtoms_singleDegree StoppableMakroAction Subpackage_CodePatterns Subpackage_JobMarket Subpackage_LinearAlgebra Subpackage_levmar Subpackage_mpqc_open Subpackage_vmg Switchable_LogView ThirdParty_MPQC_rebuilt_buildsystem TrajectoryDependenant_MaxOrder TremoloParser_IncreasedPrecision TremoloParser_MultipleTimesteps TremoloParser_setsAtomName Ubuntu_1604_changes stable
Last change on this file since 3f9eba was 3f9eba, checked in by Frederik Heber <heber@…>, 15 years ago

New classes for random number generation.

  • we use boost::random as a basis. It implements a lot of pseudo-random engines and various distributions. However, everything is templated and there are no base classes to store prototypes of the instances in a map. We need these such that we are able to offer the user at run-time every choice possible.
  • Hence, we need to encapsulate all engines, distributions and (engine, distribution) tuples: RandomNumber..._Encapsulation classes.
  • each with an abstract base class as the interface: RandomNumber... classes.
  • RandomNumberGeneratorFactory is then a singleton that contains all the Maps and tables:
    • from string (user input is string) to enum (storing enum is more efficient that a lengthy string)
    • from enum to instance (of the abstract class).
  • RandomNumberGeneratorFactory is a singleton to allow for global control of what kind of numbers are generated, hence there are also two new Actions:
  • The user in the end only uses two classes: RandomNumberGeneratorFactory and RandomNumberGenerator. The first only for obtaining a specific variant of the latter and the latter to generate the numbers (operator() is passed on to boost::variate_generator::operator()).
  • There is a unit test on the RandomNumberGeneratorFactory that checks some of the table entries and whether one of the RandomNumberGenerator is working (uniform_smallint so far).

Details:

  • As there are ~25 engines and ~15 distributions this makes an enormous amount of possible combinations. To automatically construct these, we use some stuff from boost::preprocessor in order to generate enums, tables and maps automatically just from a sequence of the typenames.
  • we have be-friended the unit test a lot to allow for easier testing, e.g. we need access to internal ..._type to test whether their type is correct.
  • cleanUp() purges the instance of RandomNumberGeneratorFactory.
  • in RandomNumber..._Encapsulation::name() functions are only used for the unit tests.
  • this all resides in its own subfolder RandomNumbers. So far, it is completely independent of everything else.
  • Property mode set to 100644
File size: 2.4 KB
RevLine 
[3f9eba]1/*
2 * RandomNumberDistribution_Encapsulation.hpp
3 *
4 * Created on: Jan 01, 2011
5 * Author: heber
6 */
7
8#ifndef RANDOMNUMBERDISTRIBUTION_ENCAPSULATION_HPP_
9#define RANDOMNUMBERDISTRIBUTION_ENCAPSULATION_HPP_
10
11// include config.h
12#ifdef HAVE_CONFIG_H
13#include <config.h>
14#endif
15
16#include <typeinfo>
17
18#include "RandomNumberDistribution.hpp"
19#include "RandomNumberGeneratorFactory.hpp"
20
21#include <boost/nondet_random.hpp>
22#include <boost/random.hpp>
23#include <boost/random/bernoulli_distribution.hpp>
24#include <boost/random/binomial_distribution.hpp>
25#include <boost/random/cauchy_distribution.hpp>
26#include <boost/random/exponential_distribution.hpp>
27#include <boost/random/gamma_distribution.hpp>
28#include <boost/random/geometric_distribution.hpp>
29#include <boost/random/linear_congruential.hpp>
30#include <boost/random/lognormal_distribution.hpp>
31#include <boost/random/normal_distribution.hpp>
32#include <boost/random/poisson_distribution.hpp>
33#include <boost/random/triangle_distribution.hpp>
34#include <boost/random/uniform_01.hpp>
35#include <boost/random/uniform_int.hpp>
36#include <boost/random/uniform_on_sphere.hpp>
37#include <boost/random/uniform_real.hpp>
38#include <boost/random/uniform_smallint.hpp>
39
40#include "unittests/RandomNumberGeneratorFactoryUnitTest.hpp"
41
42/** Template class that encapsulates the random number distributions from
43 * random::boost.
44 *
45 *
46 * We need one template parameters:
47 * -# the distribution - generates uniform random numbers
48 */
49template <class distribution>
50class RandomNumberDistribution_Encapsulation : public RandomNumberDistribution
51{
52 /**
53 * Factory is friend such that it can access private cstor.
54 */
55 friend class RandomNumberGeneratorFactory;
56 friend class RandomNumberGeneratorFactoryTest;
57
58public:
59
60 /** Getter for the type name of the internal distribution.
61 *
62 */
63 std::string name() {
64 return typeid(distribution_type).name();
65 }
66
67 /** Getter for the distribution instance.
68 *
69 * @return reference to instance
70 */
71 distribution& getDistribution() {
72 return distribution_type;
73 }
74
75 /** Constructor that instantiates a specific random number generator and
76 * distribution.
77 * @param _distribution_type instance of the desired distribution
78 */
79 RandomNumberDistribution_Encapsulation()
80 {}
81
82 /** Destructor of the class.
83 *
84 */
85 ~RandomNumberDistribution_Encapsulation() {}
86private:
87 distribution distribution_type;
88};
89
90#endif /* RANDOMNUMBERDISTRIBUTION_ENCAPSULATION_HPP_ */
Note: See TracBrowser for help on using the repository browser.