Covering a square by congruent squares
György Dósa, Zsolt Lángi, Zsolt Tuza
TL;DR
The paper determines the maximal square edge length $S(n)$ that can be covered by $n$ unit squares and the related boundary-covering value $S_{bd}(n)$. Using geometric coverings, $L$-shape embeddings, and a vertex/side/center analysis, it obtains exact small-n results, including $S(2)=1$, $S(3)=\sqrt{(1+\sqrt{5})/2}$, $S(4)=2$, and $S(5)=2$, with a parallel boundary-analysis giving $S_{bd}(4)=2$ and $S_{bd}(5)=\frac{1}{2}\sqrt{2}\sqrt{\sqrt{13-8\sqrt{2}}+1}+1\approx 2.072$, plus the recurrence $S_{bd}(n+4)=S_{bd}(n)+\sqrt{2}$ for $n\ge 4$ and a conjecture about $S(6)$. The work clarifies when boundary coverage mirrors interior coverage (for $n\le 4$) and provides precise structural descriptions of optimal configurations, including a unique boundary configuration for $n=5$. Overall, it contributes exact small-n characterizations and methodological insight into covering problems by congruent unit squares.
Abstract
The main goal of this paper is to address the following problem: given a positive integer $n$, find the largest value $S(n)$ such that a square of edge length $S(n)$ in the Euclidean plane can be covered by $n$ unit squares. We investigate also the variant in which the goal is to cover only the boundary of a square. We show that these two problems are equivalent for $n \leq 4$, but not for $n=5$. For both problems, we also present the solutions for $n=5$.
