cpitrat@fasterthanlime Wow, your approach looks very interesting, I have to dig it. My optimization came from realising that if there's only one cell not covered, it's necessarily just outside of the range covered by one sensor. So for each sensor i I just go over the positions that are at distance (d_i+1) and see if it's covered by any other (i.e. at a distance <= d_i of any sensor). This makes the approach O(nb sensor) instead if O(cells)
@fasterthanlime Ah yes indeed, your approach is simple and efficient. I feel stupid not having thought about it because I spent quite some time on this problem!