Suggestion to reduce number of run encodings/lookups/rotations needed in many cases while keeping memory use to a minimum about cubes HOT 1 OPEN

Nice suggestion!

I had the same idea, and I think it could help a lot. I would also suggest adding more potential layers of constraints to remove the possibility of degeneracy altogether. This could for example be done by calculating the center of gravity of the polycube, and orienting the cube such that it is located closest to the origo. Not sure if it fixes every case, but it should help with a lot.

Maybe the information of the projected area could be used to help reduce lookup time too. If the list of all the n=15 cubes for example is split into lists depending on the largest projected area, then any given polycube would only need to check the cubes in that given sub-list. I don't know if this would be terrible in terms of memory, but for larger size cubes it could potentially speed up the look-up time by a lot.

from cubes.