A General Lower Bound for the Limited Augmented Zarankiewicz Number based upon Complete Graphs
Combinatorics
2026-05-05 v4
Abstract
The limited augmented Zarankiewicz number zL(m,n) satisfies BSR(m,n)≥zL(m,n)≥z(m,n), where BSR(m,n) is the maximum SOS rank of m×n biquadratic forms and z(m,n) is the classical Zarankiewicz number. Our main result is a general lower bound for zL(m,n) based on the incidence graph of the complete graph K4t. For every integer t≥1, let m=(24t) and n=4t. Then zL(m,n)≥2(24t)+4t2−2t. Consequently, BSR(m,n)≥2(24t)+4t2−2t. Since z=2(24t)=Θ(t2), the gap satisfies zL−z≥4t2−2t=Θ(t2)=Θ(m), i.e., it grows linearly in m. Moreover, zzL−z≥16t2−4t4t2⟶41as t→∞, so the gap is asymptotically at least 25% of z -- a non-negligible constant fraction. For t=1 we obtain zL(6,4)≥14, and we prove that this bound is tight, i.e., zL(6,4)=14. For t=2 and t=3 we obtain zL(28,8)≥68 and zL(66,12)≥162, respectively, improving previously known bounds. We also determine the exact values of zL(m,n) for all m,n≤5: zL(5,3)=9, zL(5,4)=12, and zL(5,5)=14, confirming that previously known lower bounds are tight. These results serve as base cases for a \emph{lifting method} that constructs admissible limited augmented graphs on (m+1)×(n+1) from optimal ones on m×n. Applying this method, we obtain new lower bounds: zL(6,3)≥10,zL(6,5)≥17,zL(6,6)≥19, where z(6,3)=9, z(6,5)=14, and z(6,6)=16.
Cite
@article{arxiv.2604.04111,
title = {A General Lower Bound for the Limited Augmented Zarankiewicz Number based upon Complete Graphs},
author = {Liqun Qi and Chunfeng Cui and Yi Xu},
journal= {arXiv preprint arXiv:2604.04111},
year = {2026}
}