Darwinbots Forum

Bots and Simulations => Simulation Emporium => Topic started by: MacadamiaNuts on April 26, 2012, 11:47:58 AM

Title: Nuclear reactor zerosim
Post by: MacadamiaNuts on April 26, 2012, 11:47:58 AM
I started this week a zerosim with mutation rate fixed at 32x, maximum Point Mutations and Delta Mutations (so mutation speed can stabilize itself over time, since I'm still using an install that crashes if I change settings), costs only for DNA length, and instadeath for bots that are 10000 cycles old. I think that's the best starting settings to get a zerosim going. They started to multiply before the first 10000 cycles extinction. :P

At 3 million cycles they have developed an amusing cycle of moving a short distance, reproducing, then firing some shots. They are kinda amusing to watch.

One problem with the nuclear reactor settings is there are so many mutations the mutation history seems to cut off before 1400000 cycles. I'll have to keep more intermediate saves, I guess (automatic saving crashes too). :wacko:
Title: Re: Nuclear reactor zerosim
Post by: Botsareus on August 05, 2012, 09:55:25 AM
Already fixed the mutation history bug in the new version.

The main problem with Nuking zero bots is that:

Even though they do evolve to some what move reproduce and eat, they are too unstable to sustain this ability. I would of loved to use this method in my experiments, but I have no way to programmaticy determine when a robot has all three abilitys. It is like looking for a needle in a hay stack where the needle will eventually warp back into hay. However, you got me thinking about using this method. Did you make all the robots veg, or your starting population is over 100 robots, or both?