Equivariant bordism classification of five-dimensional \((\mathbb{Z}_2)^3\)-manifolds with isolated fixed points (Q6937070)

Let \(\mathcal{Z}_{n}((\mathbb{Z}_{2})^{k})\) be the group of equivariant unoriented bordism classes of \(n\)-dimensional \((\mathbb{Z}_{2})^{k}\)-manifolds with isolated fixed points. By classical results of Conner-Floyd and Stong, there exists a monomorphism\N\begin{align*}\N\phi_{*}:\mathcal{Z}_{*}((\mathbb{Z}_{2})^{k})\to \mathcal{R}_{*}((\mathbb{Z}_{2})^{k}),\N\end{align*}\Nwhere \(\mathcal{R}_{*}((\mathbb{Z}_{2})^{k})\) is a certain polynomial ring defined in terms of tangential representations at the isolated fixed points. This result allows one to compute \(\mathcal{Z}_{*}((\mathbb{Z}_{2})^{k})\) within the algebraic framework of the polynomial ring \(\mathcal{R}_{*}((\mathbb{Z}_{2})^{k})\). Neverthless, the ring structure of \(\mathcal{Z}_{*}((\mathbb{Z}_{2})^{k})\) for \(k\ge 3\) remains undetermined, although several partial results are known. For example, in the case \(k=3\), the structure of \(\mathcal{Z}_{n}((\mathbb{Z}_{2})^{3})\) is known for \(n = 1, 2\). For \(n = 3, 4\), the second author, together with collaborators, has obtained a complete classification: they determined the dimensions and provided explicit geometric representatives for the generators. In these cases, all generators are represented by Bott manifolds, which form a subclass of real toric manifolds.\N\NIn the paper under review, the authors investigate \(\mathcal{Z}_{5}((\mathbb{Z}_{2})^{3})\), and establish the following two main results:\N\begin{itemize}\N\item[1.] \(\dim \mathcal{Z}_{5}((\mathbb{Z}_{2})^{3})=77\).\N\item[2.] The generators of \(\mathcal{Z}_{5}((\mathbb{Z}_{2})^{3})\) can be geometrically represented by the Milnor hypersurface \(H_{2,4}\) (an \(\mathbb{R}P^{3}\)-bundle over \(\mathbb{R}P^{2}\)), and two types of 2-stage generalized real Bott manifolds: \(M_{1}\) (an \(\mathbb{R}P^{4}\)-bundle over \(\mathbb{R}P^{1}\)) and \(M_{2}\) (an \(\mathbb{R}P^{3}\)-bundle over \(\mathbb{R}P^{2}\)), equipped with certain \((\mathbb{Z}_{2})^{3}\)-actions. Notably, \(M_{1}\) and \(M_{2}\) admit extended \((\mathbb{Z}_{2})^{5}\)-actions, whereas \(H_{2,4}\) does not.\N\end{itemize}\NTo obtain these results, the authors employ a newly developed version of GKM theory, compatible with equivariant bordism, which was introduced by the second author and collaborators. Combined with the work of Conner-Floyd and Stong, their approach ultimately reduces the computation to linear algebra.