Address

Author Topic: infinite maze algorithms  (Read 3554 times)

macrutan

  • FamiLAB Member
  • Newbie
  • *
  • Posts: 35
    • View Profile
infinite maze algorithms
« on: July 20, 2012, 03:33:42 PM »
So I propose someone explore space filling curves in order to make an algorithm to auto-generate mazes. This could be done by drawing with raster or vector graphics. any takers? collaborators?


http://en.wikipedia.org/wiki/Space-filling_curve








tetsuharu

  • FamiLAB Member
  • Sr. Member
  • *
  • Posts: 487
  • My sincerest contrafibularities.
    • View Profile
Re: infinite maze algorithms
« Reply #1 on: July 21, 2012, 08:28:58 AM »
I made some giant maze printouts for Maker Faire, though I don't think any of them ever got used.

I found some python code that generated them using the Recursive division method: http://en.wikipedia.org/wiki/Maze_generation_algorithm#Recursive_division_method

What kind of program do you want to generate? Like a fractal, but a playable maze? Because I like that idea :)

johnbentcope

  • FamiLAB Member
  • Hero Member
  • *
  • Posts: 584
  • I make technology into art and art into porridge.
    • View Profile
    • JohnBentCope
Re: infinite maze algorithms
« Reply #2 on: July 21, 2012, 04:30:00 PM »
This sounds like the basis for some interesting visual elements...
I like colors.

macrutan

  • FamiLAB Member
  • Newbie
  • *
  • Posts: 35
    • View Profile
Re: infinite maze algorithms
« Reply #3 on: September 27, 2012, 12:28:38 AM »
I was throwing this out there to see what developed. Sorry I've taken so long to respond.
I'm just thinking about taking some old traditions and evolving them a step or two.


Tetsuharu nailed it. It is like a fractal (recursive), and could be an engine to generate playable mazes. One that would generate a new random maze at each initialization.  We can play with this idea this saturday if you come out for the Raspberry Pi workshop Ted is hosting. I say we start making it in python so it will run on the Pi ultimately. We can also just continue the conversation at any pace. No rush...


So for starters it would entail a maze generator that could print out kids 2d mazes on paper. Next would come a playable 2D level. The level would require navigation, collusion detection and data logging sections at the least. There are some very simple solutions for doing this that we can explore.


A third phase would take the project to a simple 3d maze. and then perhaps even further. The big idea is that we will grow our knowledge base and skills on the project. I've done this many years ago and would like to revisit the process with some colaborators. No rush, just good ol' development, production and problem solving... It would of course be open source and initiated at FamiLAB. To me it's a sort of mental puzzle  that is fun to solve and revise over time.