1. Only scored 0.2787 in Al Z's circle packing contest :)
- Posted by Al Getz <Xaxo at aol.com>
Nov 10, 2005
-
Last edited Nov 11, 2005
Hello there,
I decided to test my strategy of different diameter circle packing out by
submitting a test set in Al Zimmerman's Circle Packing Contest. I had
originally done some circle packing before and discovered a few facts
about it, but everything was done with circles ALL of the same diameter
because it was intended for a very practical use, so everything had to
be changed. The symmetry that exists with circles all of the same
size disappears when the circle radii are all different. This presented
a more or less new problem to be solved, which i think i've made progress
on.
In order to test this theory i submitted a set to be 'scored' by the
40 digit scorer used on AZ's contest site for circle packing.
To my dismay, or should i say "great dismay" (hee hee) this set only
scored 0.2787 out of a possible 1.0000 ! This is of course, wayyyy off!
Not only that, the 'scorer' tried to scale the set by something like
2.99999 or whatever when the scale i was using was 1.0000 !
So what the heck happened???
Well, <chuckle> when a set is submitted for scoring the radii must appear
in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
it makes it *easier* for the scoring program to compute the results!
Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
be a warning to others that want to try the circle packing contest:
Make sure your radii are in the correct order or you'll get a very
nasty score :)
Oh yeah, BTW, im using only Euphoria to compute the solutions, which
works out pretty nicely for up to 10 digits of accuracy (which im using
right now) because you can submit a sequence printed with 'print' if
you like such as:
{1.00,2.00},
{3.00,4.00},
...etc...
because it accepts any delimiters by simply replacing them with spaces.
Also, if anyone can figure out a practical use for packing circles
with radii ranging from 1 to N i'd sure like to hear it.
Take care,
Al
And, good luck with your Euphoria programming!
My bumper sticker: "I brake for LED's"
2. Re: Only scored 0.2787 in Al Z's circle packing contest :)
Al Getz wrote:
> I decided to test my strategy of different diameter circle packing out by
> submitting a test set in Al Zimmerman's Circle Packing Contest. I had
> originally done some circle packing before and discovered a few facts
> about it, but everything was done with circles ALL of the same diameter
> because it was intended for a very practical use, so everything had to
> be changed. The symmetry that exists with circles all of the same
> size disappears when the circle radii are all different. This presented
> a more or less new problem to be solved, which i think i've made progress
> on.
> In order to test this theory i submitted a set to be 'scored' by the
> 40 digit scorer used on AZ's contest site for circle packing.
> To my dismay, or should i say "great dismay" (hee hee) this set only
> scored 0.2787 out of a possible 1.0000 ! This is of course, wayyyy off!
> Not only that, the 'scorer' tried to scale the set by something like
> 2.99999 or whatever when the scale i was using was 1.0000 !
>
> So what the heck happened???
>
> Well, <chuckle> when a set is submitted for scoring the radii must appear
> in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
> it makes it *easier* for the scoring program to compute the results!
> Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
> be a warning to others that want to try the circle packing contest:
> Make sure your radii are in the correct order or you'll get a very
> nasty score :)
I think that should be mentioned on the contest's website!
<snip>
Did you use your "vibration mechanism"?
How about simulating a rotating lottery machine (2 dimensional, of
course)
<http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81>
Just another idea, trying to avoid thinking about the underlying
geometry. 
Regards,
Juergen
3. Re: Only scored 0.2787 in Al Z's circle packing contest :)
- Posted by Matt Lewis <matthewwalkerlewis at gmail.com>
Nov 11, 2005
-
Last edited Nov 12, 2005
Juergen Luethje wrote:
>
> Al Getz wrote:
>
> >
> > Well, <chuckle> when a set is submitted for scoring the radii must appear
> > in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
> > it makes it *easier* for the scoring program to compute the results!
> > Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
> > be a warning to others that want to try the circle packing contest:
> > Make sure your radii are in the correct order or you'll get a very
> > nasty score :)
>
> I think that should be mentioned on the contest's website!
>
It is. The instructions for submissions are quite precise.
Matt Lewis
4. Re: Only scored 0.2787 in Al Z's circle packing contest :)
- Posted by Juergen Luethje <j.lue at gmx.de>
Nov 11, 2005
-
Last edited Nov 12, 2005
Matt Lewis wrote:
> Juergen Luethje wrote:
<snip>
>> I think that should be mentioned on the contest's website!
>>
>
> It is. The instructions for submissions are quite precise.
Ooops. Because of Al's post, I believed that it is not the case,
without actually checking the website. Thanks for the correction.
Regards,
Juergen
5. Re: Only scored 0.2787 in Al Z's circle packing contest :)
Juergen Luethje wrote:
>
> Al Getz wrote:
>
> > I decided to test my strategy of different diameter circle packing out by
> > submitting a test set in Al Zimmerman's Circle Packing Contest. I had
> > originally done some circle packing before and discovered a few facts
> > about it, but everything was done with circles ALL of the same diameter
> > because it was intended for a very practical use, so everything had to
> > be changed. The symmetry that exists with circles all of the same
> > size disappears when the circle radii are all different. This presented
> > a more or less new problem to be solved, which i think i've made progress
> > on.
> > In order to test this theory i submitted a set to be 'scored' by the
> > 40 digit scorer used on AZ's contest site for circle packing.
> > To my dismay, or should i say "great dismay" (hee hee) this set only
> > scored 0.2787 out of a possible 1.0000 ! This is of course, wayyyy off!
> > Not only that, the 'scorer' tried to scale the set by something like
> > 2.99999 or whatever when the scale i was using was 1.0000 !
> >
> > So what the heck happened???
> >
> > Well, <chuckle> when a set is submitted for scoring the radii must appear
> > in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
> > it makes it *easier* for the scoring program to compute the results!
> > Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
> > be a warning to others that want to try the circle packing contest:
> > Make sure your radii are in the correct order or you'll get a very
> > nasty score :)
>
> I think that should be mentioned on the contest's website!
>
> <snip>
>
> Did you use your "vibration mechanism"?
>
> How about simulating a rotating lottery machine (2 dimensional, of
> course)
> <<a
> href="http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81">http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81</a>>
>
> Just another idea, trying to avoid thinking about the underlying
> geometry.
>
> Regards,
> Juergen
Hi there Juergen,
Yes, it actually is mentioned on the web site, i just didnt read it
carefully enough :)
Now that i think about it, if the radii were not in order you would
have to mention the radii as well as the positions, which would make
for more info needing to change hands.
Anyway, i've corrected the order and now i've got a perfect score
for my 'test' set of circles which indicates my strategy for packing
circles is all that i had hoped it would be. Sorry though, i dont want
to talk about the strategy (what was used and what wasnt used) until
after the contest is over, at which time i'll reveal and even post the
method (a single Euphoria program) used to find the solutions if anyone
happens to be interested in circle packing.
I myself wasnt very interested in circle packing at all, until i found
a very practical use for packing all circles the same diameter, at which
time i ended up with a program. Unfortunately, it doesnt work with
circles of all different diameters, however this taught me something
for this next task...if the program is general enough, it might be able
to pack any circles into any simple geometric shape...which would probably
be good enough for most practical problems in real life.
Only thing left now is...what would be a practical use for this kind of
thing? I have trouble coming up with something, except for one rather
off beat application: You'd like to display the family 'tree' as a bunch
of round pictures of different sizes of family members and you'd like to
pack them neatly into a large circle, print the whole pic out, and hang
it on the wall :)
That's about it for me so far...but im hoping i'll hear from some other
people about some more ideas for this kind of thing. When i have a
practical use for something i can feel more serious about finding a
solution.
Now it's your turn to try the circle packing contest out :)
Take care,
Al
And, good luck with your Euphoria programming!
My bumper sticker: "I brake for LED's"
6. Re: Only scored 0.2787 in Al Z's circle packing contest :)
Al Getz wrote:
> I decided to test my strategy of different diameter circle packing out by
> submitting a test set in Al Zimmerman's Circle Packing Contest. I had
> originally done some circle packing before and discovered a few facts
> about it, but everything was done with circles ALL of the same diameter
> because it was intended for a very practical use, so everything had to
> be changed. The symmetry that exists with circles all of the same
> size disappears when the circle radii are all different. This presented
> a more or less new problem to be solved, which i think i've made progress
> on.
> In order to test this theory i submitted a set to be 'scored' by the
> 40 digit scorer used on AZ's contest site for circle packing.
> To my dismay, or should i say "great dismay" (hee hee) this set only
> scored 0.2787 out of a possible 1.0000 ! This is of course, wayyyy off!
> Not only that, the 'scorer' tried to scale the set by something like
> 2.99999 or whatever when the scale i was using was 1.0000 !
>
> So what the heck happened???
>
> Well, <chuckle> when a set is submitted for scoring the radii must appear
> in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
> it makes it *easier* for the scoring program to compute the results!
> Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
> be a warning to others that want to try the circle packing contest:
> Make sure your radii are in the correct order or you'll get a very
> nasty score :)
I think that should be mentioned on the contest's website!
<snip>
Did you use your "vibration mechanism"?
How about simulating a rotating lottery machine (2 dimensional, of
course)
<http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81>
Just another idea, trying to avoid thinking about the underlying
Geometry.
Regards,
Juergen
--
Have you read a good program lately?
7. Re: Only scored 0.2787 in Al Z's circle packing contest :)
Juergen Luethje wrote:
>
> Al Getz wrote:
>
> > I decided to test my strategy of different diameter circle packing out by
> > submitting a test set in Al Zimmerman's Circle Packing Contest. I had
> > originally done some circle packing before and discovered a few facts
> > about it, but everything was done with circles ALL of the same diameter
> > because it was intended for a very practical use, so everything had to
> > be changed. The symmetry that exists with circles all of the same
> > size disappears when the circle radii are all different. This presented
> > a more or less new problem to be solved, which i think i've made progress
> > on.
> > In order to test this theory i submitted a set to be 'scored' by the
> > 40 digit scorer used on AZ's contest site for circle packing.
> > To my dismay, or should i say "great dismay" (hee hee) this set only
> > scored 0.2787 out of a possible 1.0000 ! This is of course, wayyyy off!
> > Not only that, the 'scorer' tried to scale the set by something like
> > 2.99999 or whatever when the scale i was using was 1.0000 !
> >
> > So what the heck happened???
> >
> > Well, <chuckle> when a set is submitted for scoring the radii must appear
> > in *ORDER* from 1 to N !!! Now, why this is so is beyond me, except that
> > it makes it *easier* for the scoring program to compute the results!
> > Geometrically, it doesnt matter, but to the scorer, it DOES, so let this
> > be a warning to others that want to try the circle packing contest:
> > Make sure your radii are in the correct order or you'll get a very
> > nasty score :)
>
> I think that should be mentioned on the contest's website!
>
> <snip>
>
> Did you use your "vibration mechanism"?
>
> How about simulating a rotating lottery machine (2 dimensional, of
> course)
> <<a
> href="http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81">http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81</a>>
>
> Just another idea, trying to avoid thinking about the underlying
> Geometry.
>
> Regards,
> Juergen
>
> --
> Have you read a good program lately?
>
>
Hi Juergen,
Hey thanks for reposting that link, as i almost forgot to look at it.
I think that would be pretty interesting too, but much harder because
it's in three dimensions and the physics of the machine and it's contents
would have to be simulated...neither of which are necessary in circle
packing problems (depending on your strategy of course).
I could see a program like this taking a bit of time to calculate the
next delta t positions if there are a lot of balls in the globe.
Take care,
Al
And, good luck with your Euphoria programming!
My bumper sticker: "I brake for LED's"
8. Re: Only scored 0.2787 in Al Z's circle packing contest :)
Al Getz wrote:
> Juergen Luethje wrote:
>>
>> Al Getz wrote:
>>
>>> I decided to test my strategy of different diameter circle packing out by
>>> submitting a test set in Al Zimmerman's Circle Packing Contest.
<snip>
>> How about simulating a rotating lottery machine (2 dimensional, of
>> course)
^^^^^^^^^^^^^
<http://www.stern.de/standard/popup_neu.html?id=545423&bildint=0&img=%2F_content%2F54%2F54%2F545423%2Flottofee500_500.jpg&cred=Martin+Oeser%2FDDP&width=500&height=375&bgcol=%23FFE3CF&nv=pp&navid=81>
>> Just another idea, trying to avoid thinking about the underlying
>> Geometry.
>
>
> Hi Juergen,
>
>
> Hey thanks for reposting that link, as i almost forgot to look at it.
That was thanks to Topica.
Firstly I had sent that message by mail. When I didn't see it on EUforum
after about 8 hours, I used the web interface at listfilter.com to
repost the message. Then after 1 day or so, Topica delivered the
original message ...
> I think that would be pretty interesting too, but much harder
I agree.
> because it's in three dimensions
My suggestion was to restrict it to 2 dimensions, see ^^^^^ above.
But even then it's not easy to write such a program.
> and the physics of the machine and it's contents
> would have to be simulated...neither of which are necessary in circle
> packing problems (depending on your strategy of course).
> I could see a program like this taking a bit of time to calculate the
> next delta t positions if there are a lot of balls in the globe.
In the meantime I've thought a little about the underlying geometry, and
some things are not as difficult as I had expected. 
There are still some problems that I haven't solved (yet), though.
In the other post you wrote:
| I myself wasnt very interested in circle packing at all, until i found
| a very practical use for packing all circles the same diameter, at
| which time i ended up with a program.
I understand that very well, as I'm also very often interested in the
*practical use* of things.
However, the little Euphoria program that I have so far not only
calculates the circles but also draws them on the screen, and for me the
pictures are aesthetically very pleasing.
( They remind me somewhat of the art of Victor Vasarely, which was
rather well known in the 70s ... )
In the other post you also wrote:
| Only thing left now is...what would be a practical use for this kind of
| thing? I have trouble coming up with something, except for one rather
| off beat application: You'd like to display the family 'tree' as a
| bunch of round pictures of different sizes of family members and you'd
| like to pack them neatly into a large circle, print the whole pic out,
| and hang it on the wall :)
This sounds like a good idea for a self-made Christmas present, no?
Regards,
Juergen
