#Solver matlab code

As best I can tell this will solve all solvable Sudoku puzzles, but I haven't tested any edge cases. Yes, and they were both verified as correct by an online solver afterwards. % to get around the annoying way matlab handles multiple outputs and cells % provides a list of the default sudoku pairs for 9x9 puzzleДЄssert(N=9,'only size 9 puzzle has default groupings defined') Tf = tf & all(ismember(1:size(A,1),UsedValues)) ДЇunction A = ReplaceNaNsWithCandidates(B,N) UsedValues = A(sub2ind(size(A),Groups(:,2))) % remove values from the candidates that are already taken % restriction that singular candidates cannot be repeated. % reduce the set of candidates to ones which are allowed given the PossibleSolution = BacktrackSearch(FeasibleCandidates(PossibleSolution,Groups),Groups) ДЇunction A = FeasibleCandidates(A,Groups) % pick one solution from the group of solutions % ] % note this is a complicated puzzle and does use backtrackingДЄssert(Bsz(1) = Bsz(2),'Puzzle must be square') ДЄssert(all(isnan(B(:)) | ismember(B(:),NeedNumbers)),'Only NaNs and 1:%d permitted in puzzle',N) % ] % note this is a simple puzzle and does not use backtracking % of indexes (i,j) should be considered for restriction. % B = SolveSudoku(B,Groups) allows the user to specify which groups

% B should be a 9x9 matrix constituting only NaNs, and the numbers 1-9. % and known values, produces a solution to the puzzle or errors if no % B = SolveSudoku(B): Given a Sudoku puzzle matrix B, made of NaNs This is my first try at a sudoku solver, and I just went with what seemed natural, so there may be a more efficient approach out there (in particular, it seems to be that the candidate solutions to a cell and the restrictions are duals) function B = SolveSudoku(B,Groups) I use both recursion and a semi-smart solver, and two example problems are found in the help.

#Solver matlab code

So I've created a program to solve sudoku puzzles (inspired by another question on this subreddit, found hereДЄnd I was wondering what people thought of it, and ways to improve code style or speed.