Cycling 74 » Topic: Self-avoiding walk? http://cycling74.com/forums/topic/self-avoiding-walk/feed Thu, 20 Jun 2013 01:31:03 +0000 http://bbpress.org/?v=2.2.4 http://cycling74.com/forums/topic/self-avoiding-walk/#post-65887 <![CDATA[Self-avoiding walk?]]> http://cycling74.com/forums/topic/self-avoiding-walk/#post-65887 Wed, 02 Jan 2013 23:13:42 +0000 owmtxy Has anybody played with / thought of a simple way of implementing a self-avoiding walk algorithm for working through 2D matricies?
I’m trying to ‘walk’ through a [jit.matrix] without repeating steps (which makes me think along the lines of [urn]?), but want to be able to move up/down too. Also, since the steps want to be limited to immediate neighbour cells, maybe I should be looking towards some sort of drunk/markov type thing?

Anybody got any thoughts?

O.

]]> http://cycling74.com/forums/topic/self-avoiding-walk/#post-237293 <![CDATA[Re: Self-avoiding walk?]]> http://cycling74.com/forums/topic/self-avoiding-walk/#post-237293 Thu, 03 Jan 2013 12:38:55 +0000 Roman Thilenius are you talking about a random walk in a 2d space, which does not cross its own past path?

in this case i think you wont come around creating a table with all the past coordinates, so that you can compare new steps with all the past ones.

i would probably use a signal buffer~ to do so.

-110

]]> http://cycling74.com/forums/topic/self-avoiding-walk/#post-237294 <![CDATA[Re: Self-avoiding walk?]]> http://cycling74.com/forums/topic/self-avoiding-walk/#post-237294 Thu, 03 Jan 2013 13:28:51 +0000 big_pause The trouble with a purely random walk, for anything other than a 2×2 matrix is that there is no guarantee you’ll visit every element, which is what @owmtxy wants I think. If you allow wraparound (create a torus out of your matrix), then I think you can do for 2xN, but again no guarantee for 3×3 or larger (thats taking steps to neighbours only).

A possible way to look at this is with various space filling curves.

T

]]>