C# Normal Random Number

Question

I would like to create a function that accepts Double mean, Double deviation and returns a random number with a normal distribution.

Example: if I pass in 5.00 as the mean and 2.00 as the deviation, 68% of the time I will get a number between 3.00 and 7.00

My statistics is a little weak…. Anyone have an idea how I should approach this? My implementation will be C# 2.0 but feel free to answer in your language of choice as long as the math functions are standard.

I think this might actually be what I am looking for. Any help converting this to code?

Thanks in advance for your help.

Answer 1

You might be interested in Math.NET, specifically the Numerics package.

Caveat: The numerics package targets .NET 3.5. You may need to use the Iridium package if you are targeting an earlier version...

Answer 2

This library is pretty good also:

.NET random number generators and distributions

Answer 3

Sorry I don't have any code for you but I can point you to some algorithms. See this page: http://en.wikipedia.org/wiki/Normal%5Fdistribution#Generating%5Fvalues%5Ffor%5Fnormal%5Frandom%5Fvariables

The algorithm you choose I guess depends on how accurate you want it and how fast it needs to be.

Answer 4

The MetaNumerics library, also .NET, will calculate a normal distribution (and just about anything else from statistics), super quickly. Have a look at the Feature's page for more details. The Codeplex page is here: http://metanumerics.codeplex.com/.

Answer 5

See this CodeProject article: Simple Random Number Generation. The code is very short, and it generates samples from uniform, normal, and exponential distributions.

Answer 6

Here is some C that returns two values (rand1 and rand2), just because the algorithm efficiently does so. It is the polar form of the Box-Muller transform.

void RandVal (double mean1, double sigma1, double *rand1, double mean2, double sigma2, double *rand2)
{
double u1, u2, v1, v2, s, z1, z2;

do {
    u1 = Random (0., 1.);  // a uniform random number from 0 to 1
    u2 = Random (0., 1.);
    v1 = 2.*u1 - 1.;
    v2 = 2.*u2 - 1.;
    s = v1*v1 + v2*v2;
} while (s > 1. || s==0.); 

z1 = sqrt (-2.*log(s)/s)*v1;
z2 = sqrt (-2.*log(s)/s)*v2;
*rand1 = (z1*sigma1 + mean1);
*rand2 = (z2*sigma2 + mean2);
return;

}