The paper tackles a fundamental question in the mathematics of tiling: if you can fill an infinite grid with shifted copies of several shapes simultaneously, can you always do so in a perfectly repeating, periodic pattern? This is part of a broader effort to understand when aperiodic tilings are truly necessary versus when periodicity can always be achieved. The authors work in d-dimensional integer grids and focus on collections of shapes called "tiles" that satisfy a geometric independence condition, meaning the tiles are spread out in genuinely different directions. The main result is that for d minus 1 such independent tiles in d dimensions, if any tiling exists at all, then a periodic tiling also exists. This generalizes and unifies earlier results by Bhattacharya (in two dimensions) and by Meyerovitch, Sanadhya, and Solomon (who needed an extra technical assumption the current paper eliminates).
The proof strategy involves taking a known tiling configuration and carefully transforming it into a periodic one without breaking the tiling property. The authors extend a technique called the Boolean normal-form method, originally developed by Greenfeld and Tao, which provides a structured way to decompose complicated tiling configurations into more manageable pieces. By analyzing how the tiling behaves when restricted to regular sublattices of the grid, they show that the problematic, potentially aperiodic parts of the configuration can be identified and replaced. Specifically, certain restrictions of the tiling follow a precise numerical formula involving fractional parts, and the authors show these can be replaced by constants through a careful graph-coloring argument on a finite combinatorial structure they construct.
Beyond the existence theorem, the paper delivers two practical consequences. First, it gives a complete characterization of exactly which collections of independent tiles can be periodically tiled together. Second, it provides a decision algorithm that, given such a family of tiles including an independent subset of the right size, can determine in finite time whether a joint periodic tiling exists. The dimension d is allowed to vary as an input to the algorithm, which is notable because tiling decision problems are notoriously hard or even undecidable in general. The independence condition on the tiles is what makes the problem tractable and allows the geometric structure to be fully exploited.