Evolution In Realtime (Part 2)

New mating protocal and my evolving line acquires the target AND minimises distance. WooHoo!

I had been only mating the two most fit as one group, and all the others as a separate group. Duh, no wonder it did so poorly. Using only the fittest pair left too little variation for fine tuning. All the other members were basicly blind searches. OK, Sit back and watch the perfomance of this population. The initial position was obtained after 10 minutes of fine tuning the achieved goal. I back it out of the hole and at 0:26 restart the attempt to acquire that unobtainable goal. Note how it just can not jump the bump. Near the one minute mark I change the environment/target. Evolving in the context of a static environment it gets stuck in a local minimum. Recall that I'm cyclicly varying the mutation rate. Mutation and selection is not enough to jump the bump. But there is still the outside variable, the ever changing envirnoment. It pulls the goals this way and that and over the course of time, the bump that couldn't be jumped is simply walked around. Target acquired just past the 2 minute mark.

Note the form and target position @2:43 and the change in target position @2:48 that leads to a shocking response. A seemingly stable line form that seems to be adequatly performing finds itself SUDDENLY and completely outclassed.

compare the lines at
4:23 4:40 and 4:49
Note the V shaped section that touches the corner of the purple barrier.
At 4:40 that section has been copied into a nearby section.
At 4:49 the original location for the V has become less fit and is eliminated.

This program was written in Euphoria.
Euphoria is a freely available open source interpreted language compiler.
Available at http://www.rapideuphoria.com

The source code for this gene demo is freeware developed by me and is available at
https://filer.case.edu/sxg7/public/

Part 1 is viewable here
http://www.youtube.com/watch?v=6rAe5pAxc3I

Category:

Education

Tags:

• Evolution
• simulation

[skepticoz]

So that's how you find the G spot; with an etch'a'sketch.

[damianpoirier]

@skepticoz as if you didn't know. :-)

[chaOsMastaGuru]

seems like your evolutionary pathfinder gets itself stuck at local minima. Have you (since the creation of this video) found a way of assessing fitness that doesn't result in greedy behaviour?

[damianpoirier]

wow, did you not watch the video past the one minute mark? Getting past the local minima was a huge point of the video.

[damianpoirier]

Ah, I get what you are saying now. This does not work as natural selection. All artificial selection acts, as you say, with greediness. Natural selection only removes the LEAST fit from any given generation. All other pass go and get to give their genes to the next generation. It's not hard at all to implement in an evolutionary algorithm. The main reason for not doing so is speed of execution.

[chaOsMastaGuru]

hardly, you just moved the point around to guide the thing. I watched most of the video.

[BlahKing101]

Any particular organization (Typo in the statement about peer review)

[BlahKing101]

6. There are many other factors that go into the identification of fossil species other than simply stating what the fossil looks like, but one more, when the opportunity presents itself, I'll look more into Yockey's works and see first-hand what evidence and citations he has for his claims.

You've never responded to my mentioning of endogenous retroviruses, however. I'll look into your claims, but I'd like it if you'd acknowledge mine as well.

[BlahKing101]

5. I'll check out your link later, Mmem. My personal computer's been undergoing some repairs for the past week and a half and it should be back in a few days. The one that I'm on is rather old and not entirely suitable for any manner of research. Once I get the opportunity, however, I'll further look into that site, and come to my conclusion after I've done so. (I'm not entirely familiar with Talk Origins' conduct either, so I'll check that out too)

[BlahKing101]

The scandal was the err of the particular research facility with which it was involved. The University of East Anglia was the organization at fault. Peer review is not conducted by any particular organism, rather a more or less non-structured group of experts in their respective fields freely able to view such documents at their leisure. It would be immensely difficult for the entire system to withhold the amount of data you're claiming they do.