How can get random number in your range

[bipi]

Dear all,

I found function rand(), it can create random number but this function
can not define the range of number which I want to get it, such as, I
want to get random number in the range from 0 to 100 [0-100].

Please help me,

Many thanks,

[Peter J. Holzer]

On 2007-07-27 09:38, bipi <nh*******@gmai l.comwrote:

I found function rand(), it can create random number but this function
can not define the range of number which I want to get it, such as, I
want to get random number in the range from 0 to 100 [0-100].

See the FAQ, question 13.16.

hp
--
_ | Peter J. Holzer | I know I'd be respectful of a pirate
|_|_) | Sysadmin WSR | with an emu on his shoulder.
| | | hj*@hjp.at |
__/ | http://www.hjp.at/ | -- Sam in "Freefall"

[Walter Roberson]

In article <ln************ @nuthaus.mib.or g>,
Keith Thompson <ks***@mib.orgw rote:

>Ben Bacarisse also made a good point about another problem with using
one result from rand() to generate one number is a specified range.
If there are, for example, 32768 possible results (RAND_MAX==3276 7),
and you want numbers in the range 0..99, you can't possibly produce a
uniform distribution, because 32768 is not a multiple of 100.

Or as I had already written in the random number speed thread a
couple of days ago,

"After that... find some power of 2 that is "just a bit" above a
multiple of r (minimize the difference between the power of 2
and the multiple), mask off that many bits from the returned
random number, then reject the result if it is in the upper portion
of the range (between the last full multiple and 2**N - 1)."

I didn't specifically speak of bias the way Ben did, but that
bias is the reason for the stated algorithm.
--
Okay, buzzwords only. Two syllables, tops. -- Laurie Anderson

[Keith Thompson]

ro******@ibd.nr c-cnrc.gc.ca (Walter Roberson) writes:

In article <ln************ @nuthaus.mib.or g>,
Keith Thompson <ks***@mib.orgw rote:

>>Ben Bacarisse also made a good point about another problem with using
one result from rand() to generate one number is a specified range.
If there are, for example, 32768 possible results (RAND_MAX==3276 7),
and you want numbers in the range 0..99, you can't possibly produce a
uniform distribution, because 32768 is not a multiple of 100.

Or as I had already written in the random number speed thread a
couple of days ago,

"After that... find some power of 2 that is "just a bit" above a
multiple of r (minimize the difference between the power of 2
and the multiple), mask off that many bits from the returned
random number, then reject the result if it is in the upper portion
of the range (between the last full multiple and 2**N - 1)."

I didn't specifically speak of bias the way Ben did, but that
bias is the reason for the stated algorithm.

I think for best effect (fewest rejected random numbers), you'd want
to miminize the *ratio* of the chosen power of 2 and a multiple of the
size of your range, not the difference.

Using a power of 2, I think, makes part of the operation quicker (you
can use bitwise operations rather than division), but I think that by
ignoring powers of 2 you can reject fewer of the input random numbers.

I think you can avoid rejecting *any* input random numbers at the
expense of some fairly heavy calculations. For example, if your
desired range has 100 elements, you can treat a sequence of, say, 1000
16-bit numbers as a single, very large, base-100 number, and grab one
"digit" at a time. Generalizing this to arbitrarily long sequences is
likely to require operations on arbitrarily large integer (bignums),
and will be worthwhile only if the input random numbers are
*extremely* expensive, or if you absolutely can't afford to reject too
many in a row. But for such an application, it would probably be much
easier to write your own random number generator in the first place.

If RAND_MAX+1 is a multiple of the number of elements in your desired
range, all these problems go away.

--
Keith Thompson (The_Other_Keit h) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

[websnarf]

On Jul 27, 3:07 am, "Peter J. Holzer" <[e-mail] rote:

On 2007-07-27 09:38, bipi <nh.tru...@gmai l.comwrote:

I found function rand(), it can create random number but this
function can not define the range of number which I want to get it,
such as, I want to get random number in the range from 0 to 100
[0-100].

See the FAQ, question 13.16.

The FAQ is deceptive and incomplete on this issue (and what the hell
is drand48?). You can get a better explanation here:

http://www.pobox.com/~qed/random.html

--
Paul Hsieh
http://www.pobox.com/~qed/
http://bstring.sf.net/

[Ben Bacarisse]

Keith Thompson <ks***@mib.orgw rites:

Ben Bacarisse <be********@bsb .me.ukwrites:

>Army1987 <ar******@NOSPA M.itwrites:

<snip>

>>This is already pointed out in the Web version of the FAQ.
http://www.c-faq.com/lib/randrange.html

I can't find it. I am happy to assume it is me and leave it at that.
It is not a major issue, so if it is covered somewhere, that is more
than adequate.

It's covered in the cited web page (which is question 13.16 of the FAQ):

When N is close to RAND_MAX,...

So it is. Must pay more attention.

--
Ben.

[Keith Thompson]

we******@gmail. com writes:

On Jul 27, 3:07 am, "Peter J. Holzer" <[e-mail] rote:

>On 2007-07-27 09:38, bipi <nh.tru...@gmai l.comwrote:

I found function rand(), it can create random number but this
function can not define the range of number which I want to get it,
such as, I want to get random number in the range from 0 to 100
[0-100].

See the FAQ, question 13.16.

The FAQ is deceptive and incomplete on this issue

How so?

(and what the hell
is drand48?).

That's answered in question 13.21, which question 13.16 directly
refers to.

You can get a better explanation here:

http://www.pobox.com/~qed/random.html

You raise 3 major points on that page.

1. If the desired range does not divide the range of values returned
by rand() a bias is introduced. FAQ 13.16 directly addresses that.

2. The quality of rand() is often not very good. FAQ 13.16 also
directly addresses that.

3. There are problems if the desired range is wider than RAND_MAX. I
think that's a valid criticism, though it certainly doesn't justify
using the word "deceptive" . (And in this particular context, the OP
wanted range of 0..100, and RAND_MAX is guaranteed to be at least
32767.)

Your page appears to contain a lot of good information, and I hope to
have the time to read the whole thing at some point, but it's probably
more than would be appropriate for the clc FAQ. Adding a pointer to
your page to the "additional links" page
<http://www.c-faq.com/lib/sd15.htmlmight be a good idea.

Perhaps the FAQ has been updated since the last time you read it?

--
Keith Thompson (The_Other_Keit h) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

[Ernie Wright]

we******@gmail. com wrote:

If you know anything about computer graphics; using a small state
rand() successively is like using (stochastic) dithering for images,
while using a large state rand() is like having truly higher
resolution. Having the higher resolution is generally better than
dithering which is basically trying to gloss over the fundamental
weakness of having lower resolution.

<ot>
I know something about graphics, and this analogy misses the mark.

Higher resolution with uniform samples is like a bad PRNG with lots of
bits. Stochastic sampling at lower resolution is like a good PRNG with
fewer bits. Uniform sampling produces aliasing; higher resolution just
changes the artifact wavelengths. Stochastic sampling randomizes (!)
the unavoidable error of discrete sampling, trading aliasing for much
less visually objectionable noise.

They aren't mutually exclusive, either. Just like you'd want your PRNG
with lots of bits to be good, you can stochastically sample at high
resolution. But for CG, if I have to choose, I'll take sampling quality
over raw resolution.

See

http://graphics.pixar.com/StochasticSampling/paper.pdf

So I think you need a different analogy.
</ot>

- Ernie http://home.comcast.net/~erniew