Issues with c++ 11 mersenne_twister_engine class

Question

I have been trying to use the c++ 11 mersenne_twister_engine class( http://www.cplusplus.com/reference/random/mersenne_twister_engine/) to generate numbers in the interval [0,1]

Answer 1

In C++, / on integer numbers results in an integer number. For example 11 / 2 results in 5, not 5.5. In the same way a / b is always zero if a < b.

Your problem is here:

generator()/generator.max()

generator() and generator.max() both return integers and of course, generator.max() >= generator(), so the result is zero (unless you are very lucky to get the max number in which case the result would be one).

To fix this, you can simply cast it:

(double)generator()/generator.max()

Answer 2

Perhaps you are looking for generate_canonical<double, bits>(gen)? This will generate a decimal value uniformly distributed in the interval [0,1) with bits bits of randomness (53 being the standard maximum available for a double).

Answer 3

Using one of the distributions provided in the random header probably makes more sense. As Cubbi points out std::bernoulli_distribution looks like it fits your problem well. It will generate true or false according to the distribution parameter you pass in:

#include <iostream>
#include <random>
#include <vector>
#include <algorithm>

const int WIDTH = 5 ;
const int HEIGHT = 5 ;
const double VERT_CONNECTIVITY = 0.25 ;

int main()
{
    std::random_device rd;
    std::mt19937 gen(rd());

    // give "true" 1/4 of the time
    // give "false" 3/4 of the time
    std::bernoulli_distribution d(VERT_CONNECTIVITY);

    std::vector<int> vert_con( WIDTH * HEIGHT ) ;

    std::generate( std::begin(vert_con), std::end( vert_con ), [&] () { return d(gen) ; } ) ;

    for (int i=0; i < WIDTH * HEIGHT; i++) 
    {
        std::cout << vert_con[i] << " " ;
    }
    std::cout << std::endl ;

    return 0 ;
}

Answer 4

The mersenne_twister_engine returns results of type UIntType. If you want the result to be a double, you need to cast it to a double.

double random = (double)generator()/generator.max();