Επεξεργασία

Κοινή χρήση μέσω


uniform_real_distribution Class

Generates a uniform (every value is equally probable) floating-point distribution within an output range that is inclusive-exclusive.

Syntax

template<class RealType = double>
   class uniform_real_distribution {
public:
   // types
   typedef RealType result_type;
   struct param_type;

   // constructors and reset functions
   explicit uniform_real_distribution(
      result_type a = 0.0, result_type b = 1.0);
   explicit uniform_real_distribution(const param_type& parm);
   void reset();

   // generating functions
   template <class URNG>
      result_type operator()(URNG& gen);
   template <class URNG>
      result_type operator()(URNG& gen, const param_type& parm);

   // property functions
   result_type a() const;
   result_type b() const;
   param_type param() const;
   void param(const param_type& parm);
   result_type min() const;
   result_type max() const;
};

Parameters

RealType
The floating-point result type, defaults to double. For possible types, see <random>.

Remarks

The class template describes an inclusive-exclusive distribution that produces values of a user-specified integral floating point type with a distribution so that every value is equally probable. The following table links to articles about individual members.

uniform_real_distribution
param_type|

The property member a() returns the currently stored minimum bound of the distribution, while b() returns the currently stored maximum bound. For this distribution class, these minimum and maximum values are the same as those returned by the common property functions min() and max() described in the <random> topic.

The property member param() sets or returns the param_type stored distribution parameter package.

The min() and max() member functions return the smallest possible result and largest possible result, respectively.

The reset() member function discards any cached values, so that the result of the next call to operator() does not depend on any values obtained from the engine before the call.

The operator() member functions return the next generated value based on the URNG engine, either from the current parameter package, or the specified parameter package.

For more information about distribution classes and their members, see <random>.

Example

// compile with: /EHsc /W4
#include <random>
#include <iostream>
#include <iomanip>
#include <string>
#include <map>

void test(const double a, const double b, const int s) {

    // uncomment to use a non-deterministic seed
    //    std::random_device rd;
    //    std::mt19937 gen(rd());
    std::mt19937 gen(1729);

    std::uniform_real_distribution<> distr(a,b);

    std::cout << "lower bound == " << distr.a() << std::endl;
    std::cout << "upper bound == " << distr.b() << std::endl;

    // generate the distribution as a histogram
    std::map<double, int> histogram;
    for (int i = 0; i < s; ++i) {
        ++histogram[distr(gen)];
    }

    // print results
    std::cout << "Distribution for " << s << " samples:" << std::endl;
    int counter = 0;
    for (const auto& elem : histogram) {
        std::cout << std::fixed << std::setw(11) << ++counter << ": "
            << std::setprecision(10) << elem.first << std::endl;
    }
    std::cout << std::endl;
}

int main()
{
    double a_dist = 1.0;
    double b_dist = 1.5;

    int samples = 10;

    std::cout << "Use CTRL-Z to bypass data entry and run using default values." << std::endl;
    std::cout << "Enter a floating point value for the lower bound of the distribution: ";
    std::cin >> a_dist;
    std::cout << "Enter a floating point value for the upper bound of the distribution: ";
    std::cin >> b_dist;
    std::cout << "Enter an integer value for the sample count: ";
    std::cin >> samples;

    test(a_dist, b_dist, samples);
}
Use CTRL-Z to bypass data entry and run using default values.
Enter a floating point value for the lower bound of the distribution: 0
Enter a floating point value for the upper bound of the distribution: 1
Enter an integer value for the sample count: 10
lower bound == 0
upper bound == 1
Distribution for 10 samples:
          1: 0.0288609485
          2: 0.2007994386
          3: 0.3027480117
          4: 0.4124758695
          5: 0.4413777133
          6: 0.4846447405
          7: 0.6225745916
          8: 0.6880935217
          9: 0.7541936723
         10: 0.8795716566

Requirements

Header: <random>

Namespace: std

uniform_real_distribution::uniform_real_distribution

Constructs the distribution.

explicit uniform_real_distribution(result_type a = 0.0, result_type b = 1.0);
explicit uniform_real_distribution(const param_type& parm);

Parameters

a
The lower bound for random values, inclusive.

b
The upper bound for random values, exclusive.

parm
The param_type structure used to construct the distribution.

Remarks

Precondition: a < b

The first constructor constructs an object whose stored a value holds the value a and whose stored b value holds the value b.

The second constructor constructs an object whose stored parameters are initialized from parm. You can obtain and set the current parameters of an existing distribution by calling the param() member function.

uniform_real_distribution::param_type

Stores all the parameters of the distribution.

struct param_type {
   typedef uniform_real_distribution<result_type> distribution_type;
   param_type(result_type a = 0.0, result_type b = 1.0);
   result_type a() const;
   result_type b() const;

   bool operator==(const param_type& right) const;
   bool operator!=(const param_type& right) const;
   };

Parameters

a
The lower bound for random values, inclusive.

b
The upper bound for random values, exclusive.

right
The param_type object to compare to this.

Remarks

Precondition: a < b

This structure can be passed to the distribution's class constructor at instantiation, to the param() member function to set the stored parameters of an existing distribution, and to operator() to be used in place of the stored parameters.

See also

<random>