Home > other >  Can i exclude a number or subrange of numbers inside a range of random numbers in modern C ?
Can i exclude a number or subrange of numbers inside a range of random numbers in modern C ?

Time:10-14

I have:

std::random_device rd;
std::mt19937 mt(rd());
std::uniform_int_distribution<int> probability(0, 100);

I want to exclude some numbers in this range of probabilities.
Example1: Let's say, I want to generate a random number in between 0 and 100, but this number can never be 4.
Example2: Let's say, I want to generate a random number in between 0 and 100, but this number can never be any number between 4 and 7.

I wonder if it is possible to achieve in modern C without using std::rand?
Thanks in advance!

CodePudding user response:

If you want to stay with a uniform_int_distribution you can do it manually like this:

Example1: Let's say, I want to generate a random number in between 0 and 100, but this number can never be 4.

std::random_device rd;
std::mt19937 mt(rd());
std::uniform_int_distribution<int> distribution(0,99);
auto temp = distribution(mt);
auto random_number = (temp < 4) ? temp : temp   1;

Example2: Let's say, I want to generate a random number in between 0 and 100, but this number can never be any number between 4 and 7.

std::random_device rd;
std::mt19937 mt(rd());
std::uniform_int_distribution<int> distribution(0,96);
auto temp = distribution(mt);
auto random_number = (temp < 4) ? temp : temp   4;

This could be generalize to write a function random_int_between_excluding(int first, int last, std::vector<int> exclude), though at some point it will be simpler to follow NathanOlivers suggestion and use a std::discrete_distribution instead.

CodePudding user response:

If you want to miss out 4 say, then a very good way (which doesn't compromise any statistical properties of the generator), is to draw in the half-open interval [0, 99) then add 1 if the number is 4 or greater.

You do something similar to omit numbers in a range.

This method is a surprisingly good way of modelling the quantile function associated with the desired probability distribution.

CodePudding user response:

There is an option to do it manually within a reasonable range of numbers..., create a look up table and exclude the numbers that are invalid:

 static int rand_pool[]{1,2,3,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23}; //no number 4
    srand((int)time(0));
        int random_number = rand_pool[rand() % 22];
  • Related