As can probably be gauged, I'm very bad with estimating how longs things take. The weekend plan went to hell when unexpected circumstances left me without a laptop/computer (and there's only so much you can write on paper before thinking "can I even do that?").
I have been dropped to ripping my work to shreds to find this cause, and low and behold, I have a very good idea what's causing it. The problem lies in the repulsive force calculations, it seems as though the algorithm is taking nodes of a tree which would otherwise be null, or taking its own branch, and adding the additional force.
As my coding forbid this, it's my theory that I have hacked something up elsewhere, which caused the siblings of a child to be included in the force calculation, even if they're empty quadrants of the quad tree. My code-cleaning made this a little easier to read as the hacked code is hidden from plain sight.
Whether or not this is the issue is subject to some testing (once I finish re-engineering the code), but one thing is now for sure, the problems are with the forces, specifically repulsive force (thanks to some digging, ripping and shredding).
I will update again soon, in the mean time, enjoy the wonderful cloud cover outside.
No comments:
Post a Comment