Complexity of chess domination problems
Alexis Langlois-Rémillard, Mia Müßig, Érika Róldan
TL;DR
The paper addresses the complexity of domination and maximum independent domination problems for rooks and queens on polyominoes and polycubes across dimensions. It introduces divisible gadget-based reductions from planar 3-SAT variants to prove NP-completeness for $d\ge3$, for both minimum domination and maximum independent domination, and it extends the analysis to polyominoes and convex polyominoes with conjectures. A unified ILP formulation and an accompanying solver are developed, enabling computation of new domination numbers on $n\times n$ boards up to $n=31$ and supporting a game-based exploration platform. Collectively, the work advances understanding of domination problems in higher dimensions, provides practical computational tools, and outlines open questions and conjectures for future research.
Abstract
We study different domination problems of attacking and non-attacking rooks and queens on polyominoes and polycubes of all dimensions. Our main result proves that maximum independent domination is NP-complete for non-attacking queens and for non-attacking rooks on polycubes of dimension three and higher. We also analyze these problems for polyominoes and convex polyominoes, conjecture the complexity classes, and provide a computer tool for investigation. We have also computed new values for classical queen domination problems on chessboards (square polyominoes). For our computations, we have translated the problem into an integer linear programming instance. Finally, using this computational implementation and the game engine Godot, we have developed a video game of minimum domination of queens and rooks on randomly generated polyominoes.
