Estimates of characteristic numbers of real algebraic varieties (Q817493)

A real algebraic manifold is a complex algebraic one equipped with an antiholomorphic involution. The real part \(M_\mathbb R\) of such a manifold \((M, c)\) is the fixed point set of the involution~\(c\), and the real degree of \((M, c)\) is the smallest degree of an equivariant embedding of~\(M\) in a projective space equipped with the standard real structure. According to the Harnack-Smith-Thom inequality, the total mod \(2\) Betti number of \(M_\mathbb R\) does not exceed the total mod \(2\) Betti number of~\(M\). One of the principal results of the paper under review is the following important estimate: Let~\((M, c)\) be a real algebraic manifold of dimension~\(n\) and real degree~\(d\) such that the real part \(M_\mathbb R\) of~\(M\) is orientable; then, the absolute value of any Pontryagin number of \(M_\mathbb R\) does not exceed \(2^{n^2 + 3n}d(d + n - 2)^n\). It gives in particular an explicit upper bound for the number of cobordism classes of orientable \(n\)-dimensional manifolds representable as real parts of real algebraic manifolds of a given real degree. Another result proved in the paper is the following estimate: Let~\((M, c)\) be a real algebraic manifold such that~\(M\) has complex degree~\(d_\mathbb C\); then, the total mod \(2\) Betti number \(\sum_{i = 0}^{+\infty}H_i(M; \mathbb Z/2)\) of~\(M\) (and thus, according to the Harnack-Smith-Thom inequality, the total mod \(2\) Betti number of \(M_\mathbb R\)) does not exceed \(2^{n^2 + 2}d_\mathbb C^{n + 1}\). V.~Kharlamov and F.~Zak improved the latter bound to \(\sum_{i = 0}^{+\infty}H_i(M; \mathbb Z/2) \leq d_\mathbb C^{n + 1}\) [see \textit{V. Kharlamov}, in: Algebraic geometry and geometric modelling, 103--117 (2006; Zbl 1114.14035)].