arXiv++ Combinatorics

We determine the smallest square container for eight axis-parallel squares of prescribed side lengths. The optimal side length is the minimum of the costs of 23 explicit packing strategies, and each cost is a maximum of sums of side lengths. No minimum of fewer convex functions represents it. Three strategies suffice for seven squares; the jump to 23 for eight squares witnesses the beginning of a combinatorial explosion. We prove the formula by combining the constructions with geometric lower bounds derived from disjoint coordinate projections, cornering arguments, and a weighted area comparison.

How large a square is needed to hold axis-parallel squares of prescribed side lengths? We give exact formulas for the optimal side through seven squares, with explicit constructions and geometric lower-bound proofs. For every $n\ge6$, the optimum for $n$ squares is a sum of at most $n-3$ of the prescribed side lengths. This bound is sharp on an open set of side lengths and yields the optimal Euclidean Lipschitz constant $\sqrt{n-3}$ of the optimum as a function of the $n$ side lengths.