The equivariant parametrized \(h\)-cobordism theorem, the non-manifold part (Q2118903)

This paper is part of the project of developing an equivariant Waldhausen \(K\)-theory and relating it to manifolds with group action. In [Doc. Math. 24, 815--855 (2019; Zbl 1423.19003)] the authors created a functor \(X\mapsto A_G(X)\) from \(G\)-spaces to \(G\)-spectra, where \(G\) is any finite group. The fixed-point spectrum of \(A_G(X)\) is equivalent to the \(K\)-theory of a suitable category of \(G\)-spaces that contain \(X\) as a retract, and it splits as a product \[ A_G(X)^G\sim \prod_{(H)} A(X^H_{hWH}), \] where \(H\) runs through conjugacy classes of subgroups and \(X^H_{hWH}\) is the homotopy orbit space for the action of the Weyl group on the fixed-point space \(X^H\). Of course this is reminiscent of the familiar tom Dieck splitting in equivariant stable homotopy: \[ \Sigma^\infty_G(X_+)^G\sim \prod_{(H)} \Sigma^\infty(X^H_{hWH+}). \] The main business of the present paper is to construct a natural map of \(G\)-spectra \[ \Sigma^\infty_G(X_+)\to A_G(X), \] that generalizes Waldhausen's map \(\Sigma^\infty (X_+)\to A(X)\) to the equivariant setting. The map is constructed so as to be compatible with the splittings. Thus, denoting the homotopy fiber of the map by \(\mathbf H_G^\infty(X)\), the fixed-point spectrum \(\mathbf H_G^\infty(X)^G\) splits as a product of spectra \(\mathbf H^\infty(X^H_{hWH})\). The space \(\Omega^\infty\mathbf H^\infty(Y)\) with \(Y=X^H_{hWH}\) has a description in terms of smooth manifolds: it is a homotopy colimit, over compact smooth manifolds \(M\to Y\) of increasing dimension, of the space of smooth \(h\)-cobordisms with base \(M\). This is a consequence of the Waldhausen-Jahren-Rognes work [\textit{F. Waldhausen} et al., Spaces of PL manifolds and categories of simple maps. Princeton, NJ: Princeton University Press (2013; Zbl 1309.57001)] that relates \(A(Y)\) to stabilized spaces of piecewise-linear \(h\)-cobordisms. This is an important step toward relating \(A_G(X)^G\) to spaces of high-dimensional smooth \(h\)-cobordisms with \(G\)-action. The expected next step is to consider \(\Omega^\infty \mathbf H^\infty_G(M)^G\) when \(M\) is an arbitrary compact smooth \(G\)-manifold, and to interpret it as a space of equivariant or isovariant \(h\)-cobordisms, suitably stabilized with respect to linear representations of \(G\). A few words about the technical foundations of the paper: \(G\)-spectra are considered as spectral Mackey functors on the category of finite \(G\)-sets, or equivalently as spectrally enriched functors on a certain spectral category \(\mathcal B_G\) introduced by \textit{B. Guillou} and \textit{J. P. May} [``Models of $G$-spectra as presheaves of spectra'', Preprint, \url{arXiv:1110.3571}]. It is necessary to compare \(\mathcal B_G\) with a variant called \(\mathcal B_G^{\text{Wald}}\), because the latter was used in constructing \(A_G(X)\) and obtaining the splitting of \(A_G(X)^G\). The comparison uses ``parameter multicategories''. Except for some detailed explicit constructions in this comparison argument, most of the work has been quite successfully black-boxed for readability.