SpeedOfSound wrote:Now if rocks moved then they may get in the way of the previously calculated path. The SB only has to keep trying until they get out of the way so it's still following it's original direction. If food moved then the SB could do a mini-rotation, first right, then back to true, and then left, to see where it went. Then move again.
The difference here is that we now have WALK as:
walk until food or trouble
on trouble do a random OR planned search and recalc
then proceed.
So the walk is more complex by just a bit.
Is that enough for GOB?
Can we expand on 'calculated path'?
Are you thinking in terms of some x,y position as the 'goal'?
Can we define the goal as 'feed', that might be met at many x,y positions?
If our SB gets stuck (we can define that) then take a random detour, or find another target, or something, and it might not be sound to resume the original heading or head for the original target position.
Then we can think about how the optimisation process might work to balance risk against energy gain. Wht does the SB need to know about itself to do the optimisation?