Registered: Jul 2006
Posts: 15

Hi,

Is there any interest in the "exact" (number per size +Location) of white triangles (It might works for some other structures) in the product of rule 30 (or any other rule).

Or is it just a nice exercise?
I don't want to spoil a good summer school exercise.

Martijn

Report this post to a moderator | IP: Logged

Registered: Aug 2003
Posts: 712

Certainly, there are fine questions here. Since white triangles form in rule 30 below lines of white cells, you can simply look for blocks of white n cells long. Some facts about this behavior discussed in the book can be found in notes on page 871 and page 1127.

The first makes the claim that the probability of a block of white n cells long seems to go as 2^-n, as you would expect if whites and blacks were independent and random. It also notes the first run of 10 whites occurs on step 67 from the simple initial, and length 20 first occurs at step 515.

The second note considers the step at which every block up to a given length has occurred at least once, instead, and gives the figures for lengths up to 10.

One might readily extend either question to all initials (parameterized by width, say) rather than the simplest initial condition, and push for much greater block lengths. Does the 2^-n conjecture basically hold? How long do gaps between step numbers to get every type of block get, from n to n+1? We know special initials e.g. periodic for rule 30, can and do give atypical behaviors for the rule from most or random initials. Are there identifiable classes of initials that have atypical behavior in the size of white triangles that occur in them, or are nearly all initials similar in their "white run length spectrum"?

Report this post to a moderator | IP: Logged

Registered: Jul 2006
Posts: 15

Then this might be interesting.
However it might be done one million times before...

[Edit (By Martijn)]
(I didn't really cheerfully check how and if others did this)

You don't have to check the complete triangle since every odd number of white lines in a row will make a white triangle... (So for that matter it is overly complicated.)

However if you want to count the number of occurrences of one specific structure you can use this.
[/Edit]

I'm still working on image analysis and this is just a slightly different application of some basics Ideas I use.

Maybe someone has a good use for this analyzing 2D CA?

Attachment: white_triangles.pdf
This has been downloaded 1539 time(s).

Last edited by Martijn on 03-24-2007 at 01:46 AM

Report this post to a moderator | IP: Logged

Registered: Jul 2006
Posts: 15

Just for fun, some graphs.

Attachment: example.pdf
This has been downloaded 1078 time(s).

Report this post to a moderator | IP: Logged