On submaximal curves in fake projective planes (Q2031571)

In the present article the authors give a lower bound for the multiple Seshadri \(\epsilon(X,L,r)\) constant of an ample generator \(L\) of the Picard group of a fake projective plane \(X\) with respect to \(r\) very general points \(x_1,\ldots,x_r\) of \(X\). In the case of a single point, it was proved by \textit{L. F. Di Cerbo} [J. Math. Soc. Japan 69, No. 4, 1601--1610 (2017; Zbl 1381.32011)] that the Seshadri constant \(\epsilon(X,L,1)\) is equal to Seshadri constant \(\epsilon(\mathbb{P}^2,\mathcal{O}_{\mathbb{P}^2}(1),1)\) of the hypereplane bundle on the complex projective plane with respect to a single point. On the other hand the exact multiple Seshadri constant for \(r\) very general points in a surface of Picard rank one (as the fake projective plane is) is known when \(r\) is a perfect square, so that the present article focuses in the case of \(r\neq k^2\). Combining the results of \textit{J. Roé} [J. Algebra 274, No. 2, 643--651 (2004; Zbl 1060.14008)] and \textit{L. F. Di Cerbo} [J. Math. Soc. Japan 69, No. 4, 1601--1610 (2017; Zbl 1381.32011)] it is possible to show that \(\epsilon(X,L,r)\geq\epsilon(\mathbb{P}^2,\mathcal{O}_{\mathbb{P}^2}(1),r)\). Anyway the authors show that their lower bound is better to any lower bound that one can obtain using this last inequality provided that \(10\leq r\leq 22\). On the other hand they claim that for ``\(r\) big enough (and a strong enough computer) it may be possible to find really good bounds'' for the multiple Seshadri constant.