Loading...

algogeeks@googlegroups.com

[Prev] Thread [Next]  |  [Prev] Date [Next]

[algogeeks] Re: Pbm with rand() function Don Mon Feb 27 08:01:12 2012

For instance, if RANDMAX= 32768, then

x = rand() % 20000;

is twice as likely to result in the value 10,000 as the value 15,000.
This is because there are two output values from rand() which result
in x=10000 (10000 and 30000), but only one output value from rand()
resulting in x=15000 (15000).

For any case where the statistical quality of the pseudo-random stream
is important, such as simulation, using the built-in rand() function
is not a good idea. Use a pseudo-random algorithm with much longer
period and better properties, such as Mersenne Twister.

But if you are using rand, it is usually recommended to use the high
order bits rather than the low order bits. Many implementations of
rand() have cycles in the low bits which are much shorter than the
period of the generator. He is one way to generate unbiased values of
quality as good as the generator can provide:

// Return pseudo-random integer in the range 0..n-1
int randRange(int n)
{
  int result, div = RANDMAX / n;
  do {
    result = rand() / div;
  } while(result >= n);
  return result;
}

Don

On Feb 26, 10:10 am, karthikeya s <[EMAIL PROTECTED]> wrote:
> RAND() func  returns value between 1 to INTMAX, as we know. But when
> smone tries to find out value between 1 to N he takes remainder of o/p
> of RAND() with N and adds one......but isn't it wrong coz RAND() will
> generate numbers with equal probability between 1 and INTMAX but
> taking remainder can alter the prob. of generating numbers.....
> e.g.
>
> INTMAX=50
> N=30
> RAND(50) gives numbers 1 to 30, then prob. will remain same but if it
> gives numbers 31 to 50, they'll be mapped to the numbers 1 to 20,
> which means probability of getting numbers 1 to 20 is more than the
> probability for 21 to 30.

-- 
You received this message because you are subscribed to the Google Groups 
"Algorithm Geeks" group.
To post to this group, send email to [EMAIL PROTECTED]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/algogeeks?hl=en.