Simulated evolution parlor tricks

Here are some interesting tidbits of evolutionary computing to honor Darwin’s birthday yesterday:

Evolution of Mona Lisa

(youtube link)

Roger Alsing’s idea is to start with a random pile of polygons. Random mutations are applied to the polygons. The result is compared to the Mona Lisa source image, and mutations resulting in improvements are kept. Over many generations, the evolved image begins to resemble the Mona Lisa.

This particular application of genetic algorithms is very popular. See what many other people have tried.


This site evolves music by generating loops randomly from sounds and effects. Listeners to the site’s audio streams rank the results, and the genetic algorithm creates “baby loops” for the listeners to rank.

CSS Evolve

This site shows you variations of a web site’s cascading style sheets. You pick the best results, and their genetic algorithm breeds them to create new styles for the web site.

10 responses to “Simulated evolution parlor tricks

  1. Neat! How long until we don’t need humans except to say, “That looks good.”

  2. Interesting but this isn’t really evolution as you have defined the final state condition the changes must meet in order to be “complete”. True evolution is never complete as it is system that constantly rewards positive functional changes over time in response to changes in the environment. A much better example of true simulated evolution can be found here :

    This program is a perfect example of the evolution developing characteristics un-conceived of by the author.

  3. Oh – I forgot to include the boron story as an example application of genetic algorithms:

    They used a genetic algorithm to help decode the structure of a newly discovered form of boron that is hard as diamonds.

  4. John: My first thought on that is that the “final state” is subjective. For example, “survival” as the criteria for success is very simple to define but many factors can improve survivability. In the simplistic Mona Lisa evolution, an individual “survives” only if it more closely resembles the original Mona Lisa. The final state is never quite reached, because the polygons do not seem to converge on an exact solution. Thus, it seems as though it is usually possible for some perturbation of the polygon configuration to generate an even better result despite many generations of no improvement. This compares to species of organisms that have adapted to an unchanging niche in their environment. The unchanging environment does not apply selection pressure to the organisms, and perhaps none of the emerging mutations provide any benefit to them, so the organisms remain unchanged for long periods. This also seems like a “final state” but can suddenly change after a long period.

  5. John: BTW I downloaded and ran your simulated evolution program; it is very interesting! I would love to see an updated web-based version of it. Your simulation reminds me of the entries I saw at the GECCO competitions:

  6. Pingback: Illinois Genetic Algorithms Laboratory » Evolution of Mona Lisa

  7. Pingback: Science Etcetera, Venusday 20090227 |

  8. Pingback: » Blog Archive » links for 2009-03-06: Pile o’ toys

  9. Pingback: LCS and GBML » Blog Archive » Evolution of Mona Lisa

  10. Glad you enjoyed it. It was a labor of love. Don’t see doing a web version anytime soon as there doesn’t seem to be much interest unfortunately and I have other projects going on. Regards – John

Leave a Reply