aaw wrote:In McGuire, Tugemann, and Civario's 2013 paper
There is no 16-Clue Sudoku: Solving the Sudoku Minimum Number of Clues Problem via Hitting Set Enumeration, the checker program they describe makes use of unavoidable sets to prove 16-clue puzzles don't exist. They say:
More specifically, checker contained several hundred different blueprints, where a blueprint is just a representative of an equivalence class of (minimal) unavoidable sets, the equivalence relation again being the one from Section 3.1. On the online forums, Ed Russell had investigated unavoidable sets and compiled a list of blueprints, which he sent us.
I'm trying to find these blueprints. There's a recent thread (
http://forum.enjoysudoku.com/getting-in-touch-with-red-ed-russell-t38486.html) where someone was unsuccessfully trying to track down Ed Russell -- does anyone know if he's still around these forums and/or if the list of blueprints still exists somewhere?
In the proof that no 16 clue exists, the blueprints is the basis to build the set of UAs used for each solution grid.
I started with this process (and the blueprints used by the team) when I worked on the 17 clue search.
Quickly, I switched to a process based on the brute force. It is fast, unlimited in size, and easy to implement.
As Mladen Dobrichev pushed me in this way, I suspect that gridchecker is using a similar process
Edit, in my program searching 18 clues, the UAs are mostly UAs limited to 2 bands, using a corresponding brute force
