Darwinbots Forum
Code center => Bugs and fixes => Bug reports => Topic started by: Testlund on April 15, 2008, 02:40:55 PM
-
I just found a bot that is moving anticlockwise constantly whithout any costs. Maybe it's the *aimleft store gene. First I thought it was brownian motion, but then it should have moved more random.
The bot is a small one inside a stretched out colony in the upper part of the screen, a bit to the left. Version 2.43.1i
-
The bot is turning not because it is activly turning but becuase it has angular momentum, previously aquired, probably from tie torque. Bots are not charged for passive rotation. This appears to be by design.
-
More -> Brownian motion will add angular momentum. Yours is turned up to the max with no friction to damp it. Should not be surprising that some bots are spinning in your sim...
-
Sorry for another false positive about movement costs. It's just that I haven't seen this regular rotation pattern before. I should probably have some friction setting here. Just haven't thought about adding it. What would you recommend would be the most natural for settings of Z-axis, coefficient and kinetic, if you think about the background as the bottom of a pond?
-
No worries. I'd rather have ten false positives than miss one real bug...
Personally, I like the spinning for this kind of sim. Adds unpredictability and interactivity. A bot that evolves shooting without aiming has a higher chance of hitting something for example. Set your physics how you like of course, but from a pure evo perspective, I'm a big fan of more interactivity.