Computational analysis reveals a logarithmic asymptotic bound for oblong dissections, indicating new structural invariants in squared rectangles.
Let a(n) denote the minimal number of squares in a dissection of an n × (n+1) oblong into squares (OEIS sequence A279317). We report strong numerical evidence that a(n) = (16/9) ln n + 16/9 + O(1), replacing the currently listed asymptotic b(n) = round(n^(1/3)) + 6, which we show to be valid only for n ≤ 387. The new asymptotic is supported by 15 anchor points n_k* for k = 7, …, 21, spanning n ∈ [18, 49582]. We further identify two exact combinatorial identities empirically valid on all anchor points:(i) det(L_ff) · τ(G_k) = n(n+1), where G_k is the Smith graph of the optimal dissection, τ(G_k) is its spanning-tree number, and L_ff is the doubly-reduced Laplacian.(ii) τ(G_k) = n + 1 if and only if the dissection is a simple perfect squared rectangle (SPSR, all side lengths distinct); otherwise τ(G_k) = n. The Smith graphs exhibit N(k) = ⌊(k+3)/2⌋ nodes for k ≥ 11 (with N(10) = 7 as an exceptional case), asymptotically 4-regular planar structure, and second Laplacian eigenvalue λ₂(G_k) in the range [0.98, 1.59]. The exponential rate is b_∞ = limk→∞ ln n_k* / k ∈ [0.557, 0.563], consistent with both 9/16 and ln(7/4) but indistinguishable from present data. We pose the proof of identities (i)–(ii) as open problems.
No takes yet. Share an insight, caveat, or question.
Changyuan Zeng (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: