Lower bounds on maximal determinants of binary matrices via the probabilistic method (original) (raw)

View PDF

Abstract:Let D(n)D(n)D(n) be the maximal determinant for ntimesnn \times nntimesn pm1\{\pm 1\}pm1-matrices, and mathcalR(n)=D(n)/nn/2{\mathcal R}(n) = D(n)/n^{n/2}mathcalR(n)=D(n)/nn/2 be the ratio of D(n)D(n)D(n) to the Hadamard upper bound. We give several new lower bounds on mathcalR(n){\mathcal R}(n)mathcalR(n) in terms of ddd, where n=h+dn = h+dn=h+d, hhh is the order of a Hadamard matrix, and hhh is maximal subject to hlenh \le nhlen. A relatively simple bound is \[{\mathcal R}(n) \ge \left(\frac{2}{\pi e}\right)^{d/2}
\left(1 - d^2\left(\frac{\pi}{2h}\right)^{1/2}\right)
\;\text{ for all }\; n \ge 1.\] An asymptotically sharper bound is \[{\mathcal R}(n) \ge \left(\frac{2}{\pi e}\right)^{d/2}
\exp\left(d\left(\frac{\pi}{2h}\right)^{1/2} +
\; O\left(\frac{d^{5/3}}{h^{2/3}}\right)\right).\] We also show that \[{\mathcal R}(n) \ge \left(\frac{2}{\pi e}\right)^{d/2}\] if ngen_0n \ge n_0ngen0 and n0n_0n0 is sufficiently large, the threshold n0n_0n_0 being independent of ddd, or for all nge1n\ge 1nge1 if 0ledle30 \le d \le 30ledle3 (which would follow from the Hadamard conjecture). The proofs depend on the probabilistic method, and generalise previous results that were restricted to the cases d=0d=0d=0 and d=1d=1d=1.

Submission history

From: Richard Brent [view email]
[v1] Thu, 27 Feb 2014 08:03:41 UTC (30 KB)
[v2] Thu, 13 Mar 2014 23:02:38 UTC (30 KB)
[v3] Mon, 26 Jan 2015 01:56:54 UTC (30 KB)
[v4] Sun, 1 Mar 2015 00:22:03 UTC (30 KB)
[v5] Tue, 26 Jan 2016 11:29:45 UTC (30 KB)
[v6] Tue, 25 Oct 2016 04:43:07 UTC (30 KB)