On a connection between ovoids on the hyperbolic quadric Q^+(10,q) and the Lie incidence geometry E₆,1(q)

From MaRDI portal
Publication:4385462





The hyperbolic quadric \(P={\mathcal Q}^+(2n,q)\) in \((2n-1)\)-dimensional projective space is the collection of all singular points of the quadratic form \(Q(\sum_{i=1}^{2n}\alpha_iv_i)=\sum_{i=1}^n\alpha_{2i-1}\alpha_{2i}\), where \(v_1,v_2,\ldots,v_{2n}\) is a basis for a \(2n\)-dimensional vector space over the Galois field of order \(q\). One obtains an incidence geometry of type \(D_{n,1}\) on \(P\) by taking the singular lines as the lines of the geometry. An ovoid of \(P\) is a collection of \(q^{n-1}+1\) points of \(P\) with the property that no two are orthogonal (i.e., not collinear in the induced geometry). For \(n=3\) an ovoid corresponds to an affine translation plane of dimension two over its kernel. There are a few families of ovoids known when \(n=4\). For \(n\geq 5\) it has been conjectured that ovoids do not exist. Since an ovoid exists in \({\mathcal Q}^+(2n-2,q)\) if there is one in \({\mathcal Q}^+(2n,q)\), the conjecture reduces to showing that no ovoids exist in \({\mathcal Q}^+(10,q)\). There are a number of partial results in this direction. \textit{A. Blokhuis} and \textit{G. E. Moorhouse} [J. Algebr. Comb. 4, 295-316 (1995; Zbl 0843.51011)] proved the non-existence of ovoids in \({\mathcal Q}^+(10,p^e)\) for \(p=2,3\) and \({\mathcal Q}^+(12,p^e)\) for \(p=5,7\) by using \(p\)-ranks of certain incidence matrices. Similar results were obtained by others, including the author, using different methods. The author deals with the non-existence problem by introducing a new point of view. He links \({\mathcal Q}^+(10,q)\) to the Lie incidence geometry \(E_{6,1}(q)\) as follows. Given two non-collinear points \(x\) and \(z\) in \(E_{6,1}(q)\), the geometry induced on the subspace spanned by \(\{x,z\}\cup (\Delta(x)\cap\Delta(z))\), where \(\Delta(x)\) denotes the collection of points collinear with \(x\), is isomorphic to \(D_{5,1}\). Furthermore, given a point \(x\) in \(E_{6,1}(q)\), there is a collection \({\mathcal C}\) of \(q^4+1\) such subspaces that mutually intersect only in \(x\). The author then shows that the existence of an ovoid in \({\mathcal Q}^+(10,q)\) implies the existence of a set of \(q^8+q^4+1\) pairwise non-adjacent points (co-clique) in the collinearity graph of the \(E_{6,1}(q)\) geometry. This is achieved by selecting an ovoid in each subspace in \({\mathcal C}\); the union over these ovoids yields a co-clique in \(E_{6,1}(q)\) with \(q^8+q^4+1\) points. It is hoped that the well-studied intrinsic geometry of \(E_{6,1}(q)\) and its standard embedding can be used to obtain further constraints on the ovoid in \({\mathcal Q}^+(10,q)\) and thereby eventually prove the conjecture. For the entire collection see [Zbl 0883.00017].











This page was built for publication: On a connection between ovoids on the hyperbolic quadric \({\mathfrak Q}^+(10,q)\) and the Lie incidence geometry \(E_{6,1}(q)\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4385462)