How can I generate integer random variables by using exponentially distributed with a mean of 20 seconds?

I used this code that you see in below but this all values, I generate, not integer:

mu2 = 20;

sz1 = 50;

sz2 = 1;

r2 = exprnd(mu2,sz1,sz2)

Answer 1

[Jeff Miller]

If you just want integers, then use the geometric distribution instead of the exponential:

r2 = geornd(1/mu2,sz1,sz2);

[Burhan Elaldi]

At the same time, it has to be exp. dist. I can generate with phyton code like this :

import random

for i in range(1,21)

x = int(random.expovariate(1/20))

print(x)

Can I do the same thing in the Matlab?

[Jeff Miller]

The geometric is the integer-only analog of the exponential, so it gives you exactly what you would get by truncating the exponential to an integer. Compare these two nearly identical figures:

mu2 = 20;
n=100000;
er = exprnd(mu2,n,1);
ier = floor(er);
gr = geornd(1/mu2,n,1);
edges = 0:140;
figure;
histogram(ier,edges);
figure;
histogram(gr,edges);

It really makes no sense to say that the values must be integers and must come from an exponential distribution, because the exponential distribution does not give integer values (i.e., there is zero probability that a random number from an exponential distribution is an integer).

[Burhan Elaldi]

I got it, thank you so much :)