scientific article; zbMATH DE number 1064335

From MaRDI portal
Publication:4355802





A problem in diophantine approximation which has attracted much attention is Oppenheim's conjecture on integral values of nondegenerate indefinite, irrational quadratic forms in \(n\geq 3\) variables. Extensive investigations by S. Chowla, B. J. Birch, H. Davenport, H. Heilbronn, A. Oppenheim and D. Ridout etc. resulted in verification of the conjecture for \(n\geq 21\) and partial results for a smaller number of variables. Finally, in 1986-87, Margulis proved the conjecture in full generality. The proof depended on studying dynamics of orbits of certain subgroups acting on homogeneous spaces. This method was proposed in the seventies by M. S. Raghunathan, who also proposed a conjecture about the dynamics involved which would imply Oppenheim's conjecture. Raghunathan's conjecture implies that, if \(G=\text{SL}(n,\mathbb{R})\), \(\Gamma=\text{sL}(n,\mathbb{Z})\) and \(H\) is the special orthogonal subgroup associated to a nondegenerate indefinite quadratic form \(Q\) in \(n\geq 3\) variables, any orbit of \(H\) on \(G/\Gamma\) is either closed or dense. This statement implies Oppenheim's conjecture. Margulis proved a weaker statement for \(n=3\), which along with a result of Oppenheim implies that \(Q(\mathbb{Z}^n)\) is in fact dense in \(\mathbb{R}\), and also the results for \(n\geq 3\). Subsequently the author and Margulis proved that if \(Q\) is a quadratic form as in Oppenheim's conjecture, then the set of values of \(Q\) on primitive integral \(n\) tuples is also dense. In a major development, Marina Ratner obtained a description of invariant measures of unipotent flows and used it to show that the closure of any orbit is a homogeneous subset for any Lie group \(G\) and discrete subgroup \(\Gamma\) such that \(G/\Gamma\) admits a finite \(G\)-invariant measure. She also showed that the orbits of unipotent flows are uniformly distributed in the homogeneous subsets forming their closures. Using Ratner's theorem, S. Mozes and the author were able to conclude that, given a quadratic form in \(n\geq 3\) variables as in Oppenheim's conjecture and \(\varepsilon>0\), there exists a constant \(c>0\) such that, for all large \(r\), the ball in \(\mathbb{R}^n\) with radius \(r\) and centred at \(O\) contains at least as many as \(cr^{n-2}\) integral solutions of \(|Q(x)|<\varepsilon\). Later Margulis and the author pursued the problem further and showed that the number of integral solutions of the quadratic inequalities in balls of radius \(r\) centred at \(O\) grows at least as fast as the volumes of the correspondingly defined regions in \(\mathbb{R}^n\). Finally, using similar methods, in a recent paper (preprint) A. Eskin, S. Mozes and N. A. Shah have obtained estimates on the growth of the number of solutions in large balls for certain classes of diophantine equations. For the entire collection see [Zbl 0864.00048].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4355802)