Integral Hodge conjecture for Fermat varieties

From MaRDI portal
Publication:2000271



Abstract: We describe an algorithm which verifies whether linear algebraic cycles of the Fermat variety generate the lattice of Hodge cycles. A computer implementation of this confirms the integral Hodge conjecture for quartic and quintic Fermat fourfolds. Our algorithm is based on computation of the list of elementary divisors of both the lattice of linear algebraic cycles, and the lattice of Hodge cycles written in terms of vanishing cycles, and observing that these two lists are the same.


Let \(n\) be a positive even number and \(X\subset \mathbb{P}^{n+1}\) be the Fermat variety of dimension \(n\) and degree \(d\) defined by \(\sum_{i=0}^{n+1}x_i^d=0\). Let \(\mathbf{b}=(b_0,b_1,\dots,b_{n+1})\) be a permutation of \(\{0,1,\dots,n+1\}\) such that \(b_i\) is the smallest number in \(\{0,1,\dots,n+1\}\setminus\{b_0,b_1,\dots,b_{i-1}\}\) for \(i\) an even number, and let \(\mathbf{a}=(a_1,a_3,\dots,a_{n+1})\) be a set of integers with \(0\leq a_i\leq d-1\). Let \(\mathbb{P}_{\mathbf{a},\mathbf{b}}^{n/2}\) denote the linear subvariety of \(X\) of dimension \(n/2\) defined by \(x_{b_0}-\zeta_{2d}^{1+2a_1}x_{b_1}=0, x_{b_2}-\zeta_{2d}^{1+2a_3}x_{b_3}=0, \dots, x_{b_n}-\zeta_{2d}^{1+2a_{n+1}}x_{b_{n+1}}=0\), where \(\zeta_{2d}\) is a primitive \(2d\)-th root of unity. The main result of this paper proves that the space \(V_n^d=H^n(X.\mathbb{Z})\cap H^{n/2,n/2}\) of integral Hodge cycles are generated by the classes of \(\mathbb{P}_{\mathbf{a},\mathbf{b}}^{n/2}\) when \((n,d)=(2,m)\) with \(3\leq m\leq 14\) or when \((n,d)=(4,m)\) with \(3\leq m\leq 6\). Their method of proof is based on computing the list of elementary divisors of both lattices and observing that these two lists are the same.











This page was built for publication: Integral Hodge conjecture for Fermat varieties

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2000271)