random number generator objective c

Question

Possible Duplicate:
How to generate a random float between 0 and 1?

I want to generate a random number between 0 and 1 (uniform distribution) and I use:

float x = arc4random_uniform(1);

However, this produces only 0.00000

I used

float y = arc4random() %11 * 0.1;

which returns a random number in the interval but I am not sure if it is uniform distribution.

Why isn't the first function working as expected?

You don't read the documentation, you don't accept answers, euh... — , Nov 28 '12 at 18:08
http://www.unix.com/man-page/FreeBSD/3/arc4random_uniform/ It returns an int. — Captain Giraffe, Nov 28 '12 at 18:09

score 5 · Answer 1 · 2012-11-28T23:13:59.920

5

I use:

float x = arc4random_uniform(1);

However, this produces only 0.00000

Of course it does. arc4random_uniform() returns a 32-bit unsigned integer. It does not return a float. What you're looking for is something like

#define RAND_PRECISION 1024

float x = arc4random_uniform(RAND_PRECISION) / (float)RAND_PRECISION;

Also,

I am not sure if it is uniform distribution.

Since it isn't. Using the modulo (%) operator results in a non-uniform distribution.

edited Nov 28 '12 at 23:13

answered Nov 28 '12 at 18:11

Note that there's no way to give a truely uniform distribution on a floating point interval except one roughly of the form `[N,2N-N*FLT_EPSILON]` since, in general, floating point values are not uniformly spaced. – R.. GitHub STOP HELPING ICE Nov 28 '12 at 18:32
@R.. Of course, that's why I called the divisor `PRECISION`... – Nov 28 '12 at 18:37
I did not downvote, but if your constant RAND_PRECISION was a power of two less than 2^23 (give or take a factor of 2, I am terrible at this particular kind of fencepost problem), your formula would at least give equiprobable evenly spaced values. Also, a compiler that tries to compile floating-point in a standard-conforming way has no choice but to compile `/ RAND_PRECISION` to a floating-point division. If it was a power of two, that could be optimized into a multiplication. Lastly, multiplying by `1.0f` is a strange way to convert an `int` to `float` (but the compiler should optimize it). – Pascal Cuoq Nov 28 '12 at 23:12
"I am terrible at this particular kind of fencepost problem" - I understand you. Mee too. Anyways, good suggestions and updated my answer accordingly, but that cast is really ugly. – Nov 28 '12 at 23:15
@PascalCuoq: They're not even evenly spaced unless `RAND_PRECISION` is a power of two. – R.. GitHub STOP HELPING ICE Nov 28 '12 at 23:51
@R.. Yes, that is definitely what my first sentence was intended to mean. – Pascal Cuoq Nov 29 '12 at 00:34

score 2 · Answer 2 · answered Nov 28 '12 at 18:10

2

According to the documentation, arc4random_uniform returns an integer, so using it with an upper bound of 1 won't produce what you want.

Here's the page on the arc4random functions: http://www.unix.com/man-page/FreeBSD/3/arc4random_uniform/

answered Nov 28 '12 at 18:10

RonaldBarzell

3,822
1
16
23

+1 but according to the manpage it returns an unsigned int – fvu Nov 28 '12 at 18:10
@fvu isn't an unsigned int an integer as well? Don't be nitpicky :) – Nov 28 '12 at 18:12
Thanks for the +1, but the OP expected a float, therefore unsigned or signed is irrelevant in this respect. – RonaldBarzell Nov 28 '12 at 18:12

score 0 · Answer 3 · answered Nov 28 '12 at 18:12

0

#define ARC4RANDOM_MAX      0x100000000
...
double val = ((double)arc4random() / ARC4RANDOM_MAX);

answered Nov 28 '12 at 18:12

Mohammad Rabi

1,412
2
21
41

score 0 · Answer 4 · edited May 23 '17 at 12:12

0

arc4random_uniform returns a number up to the upper bound you provide (e.g. arc4random_uniform(10) will return a number between 0 and 9). Due to that arc4random_uniform(1) should always return 0 (which is what you're seeing). If you want a random float you should see this answer:

Generate a random float between 0 and 1

edited May 23 '17 at 12:12

Community

1
1

answered Nov 28 '12 at 18:15

Mattia

2,251
1
22
27

random number generator objective c

4 Answers4