Redundant Sudoku Rules

The rules of Sudoku are often specified using twenty seven \texttt{all_different} constraints, referred to as the {\em big} \mrules. Using graphical proofs and exploratory logic programming, the following main and new result is obtained: many subsets of six of these big \mrules are redundant (i.e., they are entailed by the remaining twenty one \mrules), and six is maximal (i.e., removing more than six \mrules is not possible while maintaining equivalence). The corresponding result for binary inequality constraints, referred to as the {\em small} \mrules, is stated as a conjecture.

💡 Research Summary

The paper investigates the redundancy of the standard Sudoku constraint set, which is usually expressed as twenty‑seven “big” all_different constraints—nine for rows, nine for columns, and nine for the 3×3 blocks. By combining graphical reasoning with exhaustive exploratory logic programming, the authors identify a specific subset of six of these constraints that are logically entailed by the remaining twenty‑one. In other words, if those six constraints are omitted, every valid Sudoku solution still satisfies the full original specification. The authors prove that six is the maximal number of constraints that can be removed without altering the solution space; removing seven or more leads to at least one counter‑example where a board violates the original rules.

The methodology proceeds in three stages. First, the authors construct a bipartite graph whose vertices represent the twenty‑seven all_different constraints and whose edges connect constraints that share at least one cell. This visual model reveals that certain groups of six vertices form a dominating set: every other vertex is adjacent to at least one member of the group, suggesting a potential for logical redundancy. Second, they implement an exhaustive search in Prolog. All possible completed Sudoku grids are generated, and for each grid the satisfaction of the full constraint set is verified. The program then systematically removes constraints, testing every combination of six constraints (≈ 296,010 possibilities) to see whether the reduced set still admits exactly the same set of solutions. The search confirms that many distinct six‑constraint subsets are redundant, while any removal of seven constraints produces a grid that would be illegal under the full rule set.

Third, the graphical observations are formalized into a rigorous proof. The authors define a “redundant subgraph” and show that the six‑constraint subgraph induces a transitive closure over the remaining constraints, guaranteeing that the all_different property propagates throughout the board even without the omitted constraints. The proof relies on the fact that each all_different constraint enforces a global permutation of the digits 1‑9, and the overlap of rows, columns, and blocks creates a network of permutations that can be composed to recover the missing constraints.

In addition to the main result, the paper proposes a conjecture concerning the “small” rule set, which consists of binary inequality constraints for every pair of cells (324 constraints in total). The authors suspect that an analogous redundancy pattern exists—perhaps six binary constraints are also entailed by the rest—but they have not yet produced a formal proof. They suggest that SAT or SMT solvers could be employed to test this conjecture on larger instances.

The practical implications are notable. In constraint programming or SAT encodings of Sudoku, eliminating redundant constraints reduces the number of variables and clauses, shrinking the search space and improving solver performance. For Sudoku puzzle generators, using the minimal constraint set can speed up the validation of generated puzzles. Moreover, the combined use of graphical domination concepts and exhaustive logic programming offers a template for analyzing redundancy in other combinatorial designs such as Latin squares, KenKen, or Kakuro.

Finally, the authors outline future directions: extending the redundancy analysis to non‑standard Sudoku sizes (4×4, 16×16), experimentally verifying the impact of constraint reduction on human difficulty perception, and completing the proof of the small‑rule conjecture. The work thus contributes both a concrete theoretical insight—six is the maximal redundant subset of big Sudoku rules—and a methodological framework for studying minimal constraint systems in discrete mathematics.