On 1-2-2016 16:30, William F Pokorny wrote:
> On 02/01/2016 07:21 AM, Thomas de Groot wrote:
>> On 1-2-2016 13:19, William F Pokorny wrote:
>>> On 02/01/2016 03:06 AM, Thomas de Groot wrote:
>>>> On 1-2-2016 3:10, clipka wrote:
>>>>> Am 01.02.2016 um 02:55 schrieb Mike Horvath:
>>>>>> How do I calculate a random integer between n and p, inclusively?
>>>>>
>>>>> floor( rand(R)*((p-n)+1) + 0.5 )
>>>>>
>>>>> should do the trick
>>>>>
>>>>
>>>> Why not this: int(RRand(n,p,R))
>>>>
>>> RRand(Min,Max,RandSeededStream) from rand.inc. To get the inclusive part
>>> of the request I think we need:
>>>
>>> int(RRand(n,p+(1-1e-6),R))
>>>
>>> or
>>>
>>> floor(RRand(n,p+(1-1e-6),R))
>>>
>>> Bill P.
>>
>> Are the Min and Max not inclusive by default? I always got that
>> impression.
>>
> For floats yes.
>
> Mike wants integers so we are using int() or floor(). Seems to me the
> only way we can get the Max integer value is if the rand() function
> returns exactly 1.0 - which it will pretty much never do, but might.
>
> The +(1-1e-6) makes it so we get the Max integer value only one
> millionth less often than ideally we should.
>
> We could specify a max integer 1 larger than we really want for p, but
> then rand() will be a troublemaker and give us that 1.0 value in our scene.
>
> Bill P.
>

OK. Got it. Thanks a lot!

-- 
Thomas

Post a reply to this message

On 1/31/2016 9:08 PM, Ger wrote:
> Ger wrote:
>
>> Mike Horvath wrote:
>>
>>> How do I calculate a random integer between n and p, inclusively?
>>>
>>>
>>> Thanks
>> #declare RSeed = seed(1);
>>
> #declare RandomNumber = int(n + rand(RSeed) * (p - n));
>


Shouldn't you use floor instead of int? What if n and p are both  negative?


Mike

Post a reply to this message

On 1/31/2016 11:27 PM, Sven Littkowski wrote:
> Yes, there are two functions to convert a float to a close-by integer:
>
> 1. Floor
> If the float is, let's say, "11.4676475" (could be a multiplication of a
> "rand" random result), floor returns "11" (the next integer BELOW that
> float).
>
> 2. Ceil(ing)
> If the float is, let's say, "11.4676475" (could be a multiplication of a
> "rand" random result), floor returns "12" (the next integer ABOVE that
> float).
>


There is also int, which varies depending on if the number is greater or 
less than zero.

Post a reply to this message

On 2/1/2016 10:30 AM, William F Pokorny wrote:
> On 02/01/2016 07:21 AM, Thomas de Groot wrote:
>> On 1-2-2016 13:19, William F Pokorny wrote:
>>> On 02/01/2016 03:06 AM, Thomas de Groot wrote:
>>>> On 1-2-2016 3:10, clipka wrote:
>>>>> Am 01.02.2016 um 02:55 schrieb Mike Horvath:
>>>>>> How do I calculate a random integer between n and p, inclusively?
>>>>>
>>>>> floor( rand(R)*((p-n)+1) + 0.5 )
>>>>>
>>>>> should do the trick
>>>>>
>>>>
>>>> Why not this: int(RRand(n,p,R))
>>>>
>>> RRand(Min,Max,RandSeededStream) from rand.inc. To get the inclusive part
>>> of the request I think we need:
>>>
>>> int(RRand(n,p+(1-1e-6),R))
>>>
>>> or
>>>
>>> floor(RRand(n,p+(1-1e-6),R))
>>>
>>> Bill P.
>>
>> Are the Min and Max not inclusive by default? I always got that
>> impression.
>>
> For floats yes.
>
> Mike wants integers so we are using int() or floor(). Seems to me the
> only way we can get the Max integer value is if the rand() function
> returns exactly 1.0 - which it will pretty much never do, but might.
>
> The +(1-1e-6) makes it so we get the Max integer value only one
> millionth less often than ideally we should.
>
> We could specify a max integer 1 larger than we really want for p, but
> then rand() will be a troublemaker and give us that 1.0 value in our scene.
>
> Bill P.
>

So should I use int or floor? The result is different for numbers below 
zero.


Mike

Post a reply to this message

On 1/31/2016 10:16 PM, clipka wrote:
> Am 01.02.2016 um 03:10 schrieb clipka:
>> Am 01.02.2016 um 02:55 schrieb Mike Horvath:
>>> How do I calculate a random integer between n and p, inclusively?
>>
>> floor( rand(R)*((p-n)+1) + 0.5 )
>>
>> should do the trick
>
> ... provided you figure out that you need to add n ;)
>
> Oops.
>

Where do I add n?

Post a reply to this message

On 02/04/2016 09:49 PM, Mike Horvath wrote:
> On 2/1/2016 10:30 AM, William F Pokorny wrote:
>
> So should I use int or floor? The result is different for numbers below
> zero.
>
> Mike

Good question. If you have negative integers as part of your potential 
integer range, use floor().

Aside: In some languages to really get an integer you need int() too as 
in : int(floor(...)) because ceil() and floor() in those languages 
return floats.

Bill P.

Post a reply to this message

Am 05.02.2016 um 03:50 schrieb Mike Horvath:
> On 1/31/2016 10:16 PM, clipka wrote:
>> Am 01.02.2016 um 03:10 schrieb clipka:
>>> Am 01.02.2016 um 02:55 schrieb Mike Horvath:
>>>> How do I calculate a random integer between n and p, inclusively?
>>>
>>> floor( rand(R)*((p-n)+1) + 0.5 )
>>>
>>> should do the trick
>>
>> ... provided you figure out that you need to add n ;)
>>
>> Oops.
>>
> 
> Where do I add n?

Never mind -- I also included other nonsense in the formula.

Let me try that again:

    min( n+floor( rand(R)*((p-n)+1) ), p )

The min() is there to ensure that the result is never larger than p,
even if R is so close to 1.0 that the expression in floor() is at some
point rounded up to give a result of (p-n)+1, in which case the entire
expression (without the min()) would result in p+1.

I'm adding n only after rounding, because the lower the magnitude of a
floating point number, the higher its absolute precision.

Post a reply to this message

Am 05.02.2016 um 03:49 schrieb Mike Horvath:
> So should I use int or floor? The result is different for numbers below
> zero.

In your case you want floor(); otherwise, if your interval includes 0,
too many random numbers will be mapped to 0, and not enough random
numbers will be mapped to 0.

Post a reply to this message

On 02/05/2016 08:56 AM, clipka a dit:
> Am 05.02.2016 um 03:49 schrieb Mike Horvath:
>> So should I use int or floor? The result is different for numbers below
>> zero.
>
> In your case you want floor(); otherwise, if your interval includes 0,
> too many random numbers will be mapped to 0, and not enough random
> numbers will be mapped to 0.

*BREAKING NEWS*

Annoucing POVray 3.9, _Now_ with quantum computation !

-- 
*wip* http://la.buvette.org/vrac/essai.avi *wip*

Post a reply to this message

Am 10.02.2016 um 07:39 schrieb tth:
> On 02/05/2016 08:56 AM, clipka a dit:
>> Am 05.02.2016 um 03:49 schrieb Mike Horvath:
>>> So should I use int or floor? The result is different for numbers below
>>> zero.
>>
>> In your case you want floor(); otherwise, if your interval includes 0,
>> too many random numbers will be mapped to 0, and not enough random
>> numbers will be mapped to 0.
> 
> *BREAKING NEWS*
> 
> Annoucing POVray 3.9, _Now_ with quantum computation !

Whoops. That should read, "... and not enough random numbers will be
mapped to n."

Sorry, no support for quantum computers in the near future ;)

Post a reply to this message