I am a little confused here. Does the algorithm try to bucket nodes based on the x, y and the z axis and then finally choose the one with the lowest cost among the three? And does the buckets need to be partition space equally?