Characterising ovoidal cones by their hyperplane intersection numbers

From MaRDI portal





Given a quadric \(Q\) in a finite projective space \(P = \mathrm{PG}(d, q)\), the intersection numbers \(|Q \cap U|\) for a subspace \(U\) of the projective space \(P\) are quite restricted. For example, we have \(|Q \cap l| \in \{ 0, 1, 2, q+1 \}\) for all lines \(l\) of \(P\).\N\NAn interesting question in finite geometry is the converse: Given a set \(S\) of points of the projective space \(P\) with a small choice of possible intersection numbers, is it possible to classify the sets \(S\)? The authors study this question for subsets \(S\) of the projective space \(P = \mathrm{PG}(4, q)\) with the property that \(|S \cap R| \in \{ q+1, q^2+1, q^2+q+1 \}\) for all 3-dimensional subspaces \(R\) of \(P\). This question is motivated by the well-known fact that an ovoidal cone of \(P = \mathrm{PG}(4, q)\), i.e. a cone with vertex \(V\) a point and base an ovoid in a 3-dimensional subspace of \(P\) not containing the point \(V\) has exactly these intersection numbers.\N\NThe authors give the following interesting results:\N\NLet \(S\) be a subset of \(P = \mathrm{PG}(4, q)\) with \(|S \cap R| \in \{ q+1, q^2+1, q^2+q+1 \}\) for all 3-dimensional subspaces \(R\) of \(P\).\N\N(a) (Theorem 1.2) If \(|S \cap \pi| \geq 1\) for all planes \(\pi\) of \(P\), then \(S\) is either the point set of a plane or the point set of an ovoidal cone.\N\N(b) (Theorem 1.3 and Proposition 2.30, Remark 2.8, and Remark 2.63) Suppose that there exists a plane of \(P\) which is disjoint to the set \(S\).\N\NIf \(q > 2\), then there exists an ovoidal cone \(Y\) in \(P\) with vertex \(V\), a line \(k\) contained in \(Y\), a 3-dimensional subspace \(\Pi_k\) intersecting the cone \(Y\) at the line \(k\), and a line \(l\) of the 3-space \(\Pi_k\) intersecting the line \(k\) at a single point different from \(V\) such that \(S = (Y \setminus k) \cup l\).\N\NIf \(q = 2\), we have the above example and the following three additional examples:\N\N(i) There are three lines \(k\), \(l\), and \(m\) of \(P = \mathrm{PG}(4, 2)\) in general position (i.e. no 3-dimensional subspace of \(P\) contains all three lines \(k\), \(l\), and \(m\)) such that \(S = k \cup l \cup m\).\N\N(ii) There are two planes \(\pi_1\) and \(\pi_2\) of \(P = \mathrm{PG}(4, 2)\) which intersect in a single point \(Q\) and a hyperoval \(H\) of the plane \(\pi_2\) such that \(S = (\pi_1 \cup H) \setminus \{ Q \}\).\N\N(iii) There exists a frame \(\{ P_1, P_2, P_3, P_4, P_5, P_6 \}\) in \(P = \mathrm{PG}(4, 2)\) such that \(S = \{ P_1, P_2, P_3, P_4, P_5, P_6, Q_1, Q_2, Q_3, Q_4, Q_5 \}\) where \(Q_1, Q_2, Q_3, Q_4\), and \(Q_5\) are the third points on the lines \(P_1 P_2\), \(P_2P_3\), \(P_3 P_4\), \(P_4 P_5\), and \(P_1 P_5\), respectively.\N\NResults of this kind can often be translated into coding theory. The authors explain the corresponding result. See their Remark 1.6.











This page was built for publication: Characterising ovoidal cones by their hyperplane intersection numbers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7030814)