The Borel conjecture for hyperbolic and CAT(0)-groups (Q410111)

In the article the authors prove that closed aspherical manifolds of dimension greater or equal to \(5\) whose fundamental group belongs to a certain class \(\mathcal{B}\) of groups (specified below) are topologically rigid, i.e.~every homotopy equivalence from such a manifold to another of this sort is homotopic to a homeomorphism. The result thus provides a partial result of the so-called Borel Conjecture which claims that in fact all closed aspherical manifolds are topologically rigid. The class \(\mathcal{B}\) mentioned above is a bootstrap class of groups: it contains the hyperbolic groups as well as all groups \(G\) which act properly cocompactly and isometrically on a finite-dimensional \(CAT(0)\)-space. Moreover it is closed under taking various operations on the group, like taking subgroups, finite products, finite coproducts, and in particular it is also closed under taking arbitrary directed colimits. Specifically the last property ensures that the class \(\mathcal{B}\) is quite big. The two authors show the result by verifying the so-called Farrell-Jones Conjecture in \(K\)- and \(L\)-theory for all the groups in the class \(\mathcal{B}\) in low resp.~all degrees, and indeed \textit{C. Wegner} in [Proc. Am. Math. Soc. 140, No. 3, 779--793 (2012; Zbl 1240.19003)] showed that the assertion of the Farrell-Jones Conjecture in \(K\)-theory also holds in the remaining degrees for groups in the class \(\mathcal{B}\). Besides its relevance for the Borel Conjecture the verification of the Farrell-Jones Conjecture in \(K\)- and \(L\)-theory has a lot of additional applications like the Bass Conjecture, the Kaplanski Conjecture, or the Novikov Conjecture; for an overview e.g.~see [\textit{W. Lück} and \textit{H. Reich}, ``The Baum-Connes and the Farrell-Jones conjectures in \(K\)- and \(L\)-theory'', in: E. M. Friedlander (ed.) et al., Handbook of \(K\)-theory. Vol. 1 and 2. Berlin: Springer. 703--842 (2005; Zbl 1120.19001)]. The main strategy for the proof is similar to the one used in [\textit{A. Bartels}, \textit{W. Lück} and \textit{H. Reich}, Invent. Math. 172, No. 1, 29--70 (2008; Zbl 1143.19003)] where the authors use geometrically controlled algebra in order to prove the Farrell-Jones Conjecture in \(K\)-theory for hyperbolic groups. Besides establishing the inheritance properties of the class \(\mathcal{B}\) two main new results in the article under review consist of an adaptation of the arguments in [Bartels, Lück and Reich, loc. cit.] to the case of groups which act properly cocompactly and isometrically on a finite-dimensional \(CAT(0)\)-space, as well as the extension of the methods and ideas from the \(K\)-theory context to the \(L\)-theory context. Here is a very rough description of how the major results are obtained: like in [Bartels, Lück and Reich, loc. cit.], the authors first translate the statement of having an isomorphism in \(K\)- and \(L\)-theory into a vanishing result for the \(K\)- resp. \(L\)-theory of an associated obstruction category and then consider corresponding representing geometrically controlled cycles for the \(K\)-theory of the obstruction theory. After verifying a version of the so-called stability theorem for this context (which corresponds to Theorem 5.3. in the paper under review) it suffices to show that every \(K\)- resp. \(L\)-theory class for the obstruction category is represented by a geometrically controlled cycle, where the associated geometric control (which is a non-negative real number) is smaller than an a priori given (typically very small) positive value \(\varepsilon\). And it then follows that the cycle at hand indeed represents the zero element in the \(K\)-theory of the obstruction theory. For proving the result in the \(K\)-theory context the authors then use a transfer map to alter the underlying parameter space where the geometric control takes place so that the new parameter space comes equipped with an additional useful geometric structure. In the case of hyperbolic groups the corresponding space is an analogue of the tangent sphere bundle of the universal covering of a closed hyperbolic manifold, and the additional structure on it corresponds to an analogue of the geodesic flow; in the case where \(G\) acts appropriately on a \(CAT(0)\)-space the authors use as a replacement an associated flow space introduced in [\textit{A. Bartels} and \textit{W. Lück}, Geom. Topol. 16, No. 3, 1345--1391 (2012; Zbl 1263.37052)] which has similar formal properties. In both cases they can use the flow to change an arbitrarily given cycle for the \(K\)-theory of the obstruction category within its \(K\)-theory class and thereby diminish the geometric control in such a way that they can use the stability theorem above to derive that the cycle actually represents the zero element. Here one technical difficulty lies in the fact that the group in the \(CAT(0)\)-case only acts up to homotopy on the associated flow space, and it is one of the major achievements of the authors that they manage to overcome this specific difficulty. When adapting the strategy of proof from the \(K\)-theory to the \(L\)-theory context the authors were forced to use some different transfer map for the argument. In particular the properties of the newly adapted parameter space are now much more subtle and cause further complications. And as mentioned above, the fact that the authors nevertheless succeed in carrying over the whole strategy from the \(K\)-theory proof can be regarded as a second major achievement documented in this article.