I decidedly LOVE riding on a train: For whatever reason, it’s one particular moment where I somehow focus a bit more than usual. It’s usually just a few hours, reasonably quiet environment, a bit more spacious than say airplanes (airplanes are just not the same, even with earplugs), no TV to distract my attention, the excuse of limited/imperfect mobile coverage… You name it. I just don’t know, but it works, I feel motivated to code & run experiments while on a train…
Introducing the Cellular Automata
One thing that I liked about the book I recommended a couple of days back (see last entry before this one) is that it provides nice and easy examples (but in Python), and then reasonable exercises. I loved to read through most of them (and some of the math sections), but reading is not that fun in this case, I wanted to try it out (of course!).
The following is the result of a simple R implementation of “Cellular Automata” for simulation of a “fire spread” in a theoretical forest setting. The identification of the “neighbour” trees in fire follows the Von Neumann definition of “neighbours” in this 2D configuration.
Hopefully the video is quite self-explanatory. You start with ONE burning tree, and then let time pass… 🙂
Visualization trick
One thing I learnt today is how to quickly draw a matrix into a picture. I hereby recommend you look into the “MBCbook” R package, for its function “imshow()”. Which incidentally I found while looking for alternatives to the Python function… imshow! (So yeah, it was a fast search that one…)
Conclusions
I hope you enjoyed it (I could definitely make it faster, and more or less dense a forest, and change the neighbours identification, and what not… I’ve tested this a few times with a few parameters).
At least to me, this was for no-good-reason quite… Satisfying 🙂
And I’m not sharing the code (not yet anyway) just because this was a first, horrible pasted-together step by step implementation, full of slow and nested “for loops”, to-be-improved matrix indexing, not enough functions… And well, just not quite “presentable”. Find the code linked below. But it does work just fine though 🙂