21–30 of 75 posts
This looks like fun! It seems that the copied champions are simulated in every run. As long there is no random influence you can probably skip this simulation, as the simulation of the champions is also most likely the most CPU intensive. EDIT: Grammar.
Fun, but even after hundreds of generations the walkers are still pretty bad. How good would they be after thousands or millions of generations? Is it even feasible to create a decent walker using genetic algorithms?
What exactly is evolving? How is the genome represented, how does it control the walking?
The rotation speed for each joint on each simulation step is given by:
x * cos(y + z * simulation_steps);
I was just about to give up at 70th generation, when the walkers suddenly evolved from 1 to 4 steps in 5 generations. Currently at generation 120 with 7 steps, looks like it is "evolving" in bursts, with long periods of nothing. Well it is stuck at 9 steps at gen 400+. 327 aibobe baeado 948.93 >> 548 aibobe baeado 1058.76 I wonder if the author got any further. I would say this is close to the limit.
What exactly is evolving? How is the genome represented, how does it control the walking?
The genome is represented by a set of parameters that control the rotation speed on each joint, 3 parameters per joint. The rotation speed for each joint on each simulation step is given by: x * cos(y + z * simulation_steps); where x, y, and z are the parameters. I've experimented with more sophisticated models, but it was hard to get any kind of evolution in the attention span people give to a browser game. :D
This is also an important feature in the SIMBICON controller, which is arguably the simplest and most robust walker system. (http://www.cs.ubc.ca/~van/papers/Simbicon.htm)
EDIT: and yes, Reinforcement Learning is much more effective way of attacking this type of problem (e.g. see some recent work from Sergey Levine http://www.eecs.berkeley.edu/~svlevine/), but I also agree with the author that GA are fun :)
Fun, but even after hundreds of generations the walkers are still pretty bad. How good would they be after thousands or millions of generations? Is it even feasible to create a decent walker using genetic algorithms?
It doesn't really say how it works, but it doesn't seem like a very natural way to do walking. E.g. here is are evolved walkers in a more complicated 3d simulation: http://vimeo.com/79098420 They seem to get very good after just a few generations compared to this.
Fun, but even after hundreds of generations the walkers are still pretty bad. How good would they be after thousands or millions of generations? Is it even feasible to create a decent walker using genetic algorithms?
This experiment reminds me of one of my recent experiments: http://madebyevan.com/nn-gait/ . Mine ends up figuring how to walk relatively quickly. It does have a regular pulse as an input though.
Earlier quoted context omitted.
This experiment reminds me of one of my recent experiments: http://madebyevan.com/nn-gait/ . Mine ends up figuring how to walk relatively quickly. It does have a regular pulse as an input though.
Quite interesting! Is there any source code available publicly for this?