Bots and Simulations > Mutations
Swimy(F1)(Darwinbots)4-11-2016
Botsareus:
I ended up reverting the fittest code as you suggested. It really has little impact on how much the robots seem to lose in survival mode and if anything fittest can be configured to only favor population so that solves that little problem. However, I am still keeping my fitness function based on energy for now. I might reverting it as well as a last resort. Lets see what happens and sorry for the long wait between version drops if I had more computing power it would go allot faster.
spike43884:
--- Quote from: Botsareus on April 27, 2016, 12:38:53 PM ---I ended up reverting the fittest code as you suggested. It really has little impact on how much the robots seem to lose in survival mode and if anything fittest can be configured to only favor population so that solves that little problem. However, I am still keeping my fitness function based on energy for now. I might reverting it as well as a last resort. Lets see what happens and sorry for the long wait between version drops if I had more computing power it would go allot faster.
--- End quote ---
Ok. Np.
OFFTOPIC: I'm not sure if I mentioned this in a post, but I was thinking, we could try moving to a new, shinier forum (there has to be some fancier one out there now), and then It'll be lovely and clean for any newcomers. This forum could then be a sort of archive.
Botsareus:
As expected it is working a bit worse even with the effect of body reduced by 3/10. I am not taking body out completely but I might end up reducing it to 1/10. I can also try the other way and revert both the fittest and the survival fitness to normal body.
spike43884:
--- Quote from: Botsareus on April 27, 2016, 02:54:56 PM ---As expected it is working a bit worse even with the effect of body reduced by 3/10. I am not taking body out completely but I might end up reducing it to 1/10. I can also try the other way and revert both the fittest and the survival fitness to normal body.
--- End quote ---
What about factoring in kills per bot.
So the fittest function records the kills each bot has, till it dies, which that becomes a fixed figure, the average tells you how successful a species is, for individual bots look at their individual kills?
Botsareus:
Damn, just tried a zerobot run with my better? mutation algorithms robots are just sitting there without doing anything for a long time. I am seriously considering reverting most of my changes and just keeping the bug fixes.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version