Darwinian Natural Selection

For a long while I've wanted to see a time-lapse view of how an image evolves. I imagined there were dynamics and nuances that were impossible to see watching evolution second-to-second. So after writing my last post (and with a small nudge from my friend Daniel), I set the evolver to output an image after each 1% …

Artistic Deconstruction

After playing about with animations of evolved lines, I took a step back to do something similar with the original triangle images. I've recently refined the algorithm to improve the results. Two different tweaks have improved the evolutionary process (a little intelligent design if you will). First, when adding new triangles I colour each vertex …

Avengers, Disassemble! (Then totally reassemble!)

In the last update I applied the ideas of Petros Vrellis to my image evolver, to recreate pieces of art out of a single piece of thread. If occurred to me that it would be quite easy to animate between different evolved images, if they contained the same amount of lines. I could evolve one image …

The single threaded loom

As I was wrapping up my triangular image evolution experiments, I was inspired by a couple of posts by my friend Mike McCarthy, a concept artist and fellow Fable veteran. The first post that piqued my interest was this dread-inducing Spider Queen he'd just finished colouring. Certainly the stuff of nightmares and a testament to his skills as …

An enigma in triangles

After finishing, tidying up and checking in my code for evolving minimal sorting networks, I decided to move on to the more aesthetically pleasing challenge of evolving images out of triangles. This experiment was first carried out back in 2008 by Roger Johansson. You can read his blog post and results in full, but the essence is that …

So, what is a genetic algorithm?

To herein be found my ongoing, annotated, plausibly correct description of a genetic algorithm... A genetic algorithm is a process that models the biological process of evolution by natural selection: organisms that manage to successfully breed before they shuffle off this mortal coil being considered tip-top. Those that don't not really being considered at all. …

A First Stab at Evolution

I had two goals in reimplementing Hillis' minimal sorting networks experiment: to start carving a generalised framework for evolutionary experiments and to try and improve on my previous attempts from university to find optimal sorting networks. While it would be all well and good* to sit down and try to construct a detailed design for an evolutionary …

Minimal Sorting Networks

Back in 2000 I went back to university. A friend of mine who was just about to start his PhD at Sussex Uni told me about a master's degree called Evolutionary and Adaptive Systems. The courses sounded like all the bits of my degree that I'd enjoyed the most (evolution, simulated robotics, neural networks) and none of …

Out of the Primordial Soup

Not long ago* I started out on a bold new coding adventure to create a toolkit for performing simulated evolutionary experiments. I have a history of such things, having played around with evolution for 20 years or so, starting with little experiments at university consisting of neural networks and simulated robots in simple little artificial worlds. Evolution is fascinating for …