Am 2017-06-23 06:09, also sprach Stephen:

> I've spent two days making rocks and trying to lay them in a single 
> layer. I'm really struggling with it.
> 

I thought about it a little. I can easily do this in an animation, 
because of the constant recursive adjustments and tweaking of positions. 
  Building a single static image with everyone in the right place, not 
so much.

-- 
dik

Post a reply to this message

Thomas de Groot <tho### [at] degrootorg> wrote:

> > I will just say that there apparently has been much work done on exactly this.

Whoops - by "this", I meant the original objective.

I suppose that's the old / whole problem with pronouns, and I didn't want to
just hand out the links I was perusing - just yet.  ;)

I figured we'd see how Ari was coming along on this.

Post a reply to this message

On 26-6-2017 13:38, Bald Eagle wrote:
> Thomas de Groot <tho### [at] degrootorg> wrote:
> 
>>> I will just say that there apparently has been much work done on exactly this.
> 
> Whoops - by "this", I meant the original objective.
> 
> I suppose that's the old / whole problem with pronouns, and I didn't want to
> just hand out the links I was perusing - just yet.  ;)
> 
> I figured we'd see how Ari was coming along on this.
> 
> 
> 

I'll sit back and watch :-)


-- 
Thomas

Post a reply to this message

http://www.afhalifax.ca/magazine/wp-content/sciences/EmpilementDeDisques/ProducingPacking/Initialization.pdf

and

https://arxiv.org/pdf/1005.0011.pdf

 "For generally
shaped particles, finding the densest packings is notoriously difficult. This
salient point is

we
cannot guess or even closely approximate the answer, let alone prove it, and it
is difficult to
develop even a qualitative understanding of the effects of grain shape on

Until recently, very little was known about the densest packings of polyhedral
particles. The
difficulty in obtaining dense packings of polyhedra is related to their complex
rotational
degrees of freedom and to the non-smooth nature of their shapes [28, 29]."

Post a reply to this message

Amidst the chaos of RL, (I moved over the weekend) I managed to work out a
reliable placement of a sphere tangent to all 4 faces of any irregular
tetrahedron.
I worked out some macros to generate the determinant of a 4x4 matrix, based on
the Laplace expansion, or expansion by minors, but for whatever reason that
didn't play nicely with my tetrahedra.
I might have to shovel a lot of output to the debug stream and work out a smaple
problem by hand to see if there's a bug somewhere.

I modified some code from John Burkardt, and that seemed to work, so ...


It's immediately apparent that there will be large gaps between such spheres in
a random tetrahedral "mesh", and these guys have done some work on that next
step:
http://www.afhalifax.ca/magazine/wp-content/sciences/EmpilementDeDisques/ProducingPacking/Initialization.pdf

certainly one can see how the Delaunay-optimized triangles make the incircles
pack much more nicely, so that's a fairly important first-step to filling space
by this method.

Wondering how Ari's doing....    :D

Post a reply to this message

"Bald Eagle" <cre### [at] netscapenet> wrote:
> Amidst the chaos of RL, (I moved over the weekend) I managed to work out a
> reliable placement of a sphere tangent to all 4 faces of any irregular
> tetrahedron.
> I worked out some macros to generate the determinant of a 4x4 matrix, based on
> the Laplace expansion, or expansion by minors, but for whatever reason that
> didn't play nicely with my tetrahedra.
> I might have to shovel a lot of output to the debug stream and work out a smaple
> problem by hand to see if there's a bug somewhere.
>
> I modified some code from John Burkardt, and that seemed to work, so ...
>
>
> It's immediately apparent that there will be large gaps between such spheres in
> a random tetrahedral "mesh", and these guys have done some work on that next
> step:
>
http://www.afhalifax.ca/magazine/wp-content/sciences/EmpilementDeDisques/ProducingPacking/Initialization.pdf
>
> certainly one can see how the Delaunay-optimized triangles make the incircles
> pack much more nicely, so that's a fairly important first-step to filling space
> by this method.
>
> Wondering how Ari's doing....    :D

Hi, guys! Haven't been here for a while because I am trying to find a way which
I could 'understand'.(I have found a lot of research papers on this topic but
none of them are comprehensible for me) Therefore, I ended up finish my project
*in a very simple way*. Just make random spheres and test with distance
equation. And that's it. :(
http://i.imgur.com/xH64ud9.jpg
*Dawn it! I hope I can find a way to improve the code so that they won't float*

Post a reply to this message

On 7/25/2017 7:36 PM, Ari wrote:
>> Wondering how Ari's doing....    :D
> Hi, guys! Haven't been here for a while because I am trying to find a way which
> I could 'understand'.(I have found a lot of research papers on this topic but
> none of them are comprehensible for me) Therefore, I ended up finish my project
> *in a very simple way*. Just make random spheres and test with distance
> equation. And that's it.:(
> http://i.imgur.com/xH64ud9.jpg
> *Dawn it! I hope I can find a way to improve the code so that they won't
> float*
>
>


Good of you to come back. :)

I said at the beginning that it was not trivial. I wonder what your 
lecturer's solution would be?

Your image looks a good solution but you are an Engineer. What were your 
tolerances?

Now you know your way here. Don't be a stranger. :)



-- 

Regards
     Stephen

Post a reply to this message

> > Hi, guys! Haven't been here for a while because I am trying to find a way which
> > I could 'understand'.(I have found a lot of research papers on this topic but
> > none of them are comprehensible for me)

*** Yet ***    ;)

> Therefore, I ended up finish my project
> > *in a very simple way*. Just make random spheres and test with distance
> > equation. And that's it.:(
> > http://i.imgur.com/xH64ud9.jpg
> > *Dawn it! I hope I can find a way to improve the code so that they won't
> > float*

Excellent.  It's important to stick with it and DO _something_, even if it's not
ideal.  Never let the perfect be the enemy of the good.
I like your results.
Now you know more about how "the devil is in the DETAILS", and it's NEVER _just_
"All you have to do is..."

As a computer graphics trick, why don't you try increasing the size of all those
rocks by 10, and then take the whole thing and scale it by 1/10.
The apparent size of the rocks might stay the same, while the spaces between
them that make them look like they're floating (because they are) might be
reduced by 9/10.

If that works, try doing the same with 100 or 1000 instead of 10.


> I said at the beginning that it was not trivial. I wonder what your
> lecturer's solution would be?

If it's an engineering class, he might not have one - it might be an open-ended
exercise to see what people come up with.


> Now you know your way here. Don't be a stranger. :)

Indeed, there's always more fun to be had, and so very many things to learn.

Check out:
https://malbertsinvention.wordpress.com/2013/08/15/realistic-dry-stone-wall-physics-test/
https://www.google.com/search?q=rocksolver
https://www.google.com/search?q=robot+arm+stack+stones

Post a reply to this message

> Good of you to come back. :)
>
> I said at the beginning that it was not trivial. I wonder what your
> lecturer's solution would be?
There is no solution :( The professor just wanted to let us try something *new*
and see if we are interested in this software. Overall, I think this software is
pretty fun. I think I will come back after attending more math class in the
University :D.
>
> Your image looks a good solution but you are an Engineer. What were your
> tolerances?
>
Same as the above. We don't have any 'standard' or 'marking scheme'. My
professor just found out that amazing grains picture and he told us it could be
a great example for us to begin with and imitate. So, I need to figure out how
to render a similar scene :D.
> Now you know your way here. Don't be a stranger. :)
Ayyyyy, thank you Stephen B-).

Post a reply to this message

> Excellent.  It's important to stick with it and DO _something_, even if it's not
> ideal.  Never let the perfect be the enemy of the good.
That's true! Trying to be *perfect* is actually the culprit of my anxiety. After
I changed from *perfect mode* to *good mode*, I started to appreciate this
software. What a great invention :D.
> I like your results.
> Now you know more about how "the devil is in the DETAILS", and it's NEVER _just_
> "All you have to do is..."
>
> As a computer graphics trick, why don't you try increasing the size of all those
> rocks by 10, and then take the whole thing and scale it by 1/10.
> The apparent size of the rocks might stay the same, while the spaces between
> them that make them look like they're floating (because they are) might be
> reduced by 9/10.
That's cute :P. I think I will try it later.
> If that works, try doing the same with 100 or 1000 instead of 10.
>
>
> > I said at the beginning that it was not trivial. I wonder what your
> > lecturer's solution would be?
>
> If it's an engineering class, he might not have one - it might be an open-ended
> exercise to see what people come up with.
Yes, it is. :D
>
> > Now you know your way here. Don't be a stranger. :)
>
> Indeed, there's always more fun to be had, and so very many things to learn.
>
> Check out:
>
https://malbertsinvention.wordpress.com/2013/08/15/realistic-dry-stone-wall-physics-test/
> https://www.google.com/search?q=rocksolver
> https://www.google.com/search?q=robot+arm+stack+stones
They are some interesting examples, especially the robot one. The Rocksolver is
great, too! Thank you so much B-).

Post a reply to this message