Algebraic \(K\)-theory over the infinite dihedral group: an algebraic approach (Q640318)

The infinite dihedral groups has two descriptions. The first is as the free product of two cyclic groups of order \(2\). The second is as the non-trivial semidirect product of the infinite cyclic group by the cyclic group of order \(2\). The second description implies that the infinite dihedral group has an infinite cyclic subgroup of index \(2\). So if a group \(G\) admits an epimorphism to the infinite dihedral group, first of all, it admits a splitting as an amalgamated free product. Furthermore, it contains a subgroup \(\bar{G}\) of index \(2\), that is a semidirect product of a group by the infinite cyclic group. One direct summand of the algebraic \(K\)-theory of \(RG\) (\(R\) a ring) is the Waldhausen's Nil-groups, the other is its controlled part over the interval. There is a similar splitting for the \(K\)-theory of \(R\bar{G}\). In this case the Nil-groups split also into two summands, each summand being a Bass-Farrell-Hsiang Nil-group. In this paper, the authors prove that the Nil-groups for \(RG\) are isomorphic to one of the Nil-groups of \(R\bar{G}\) (the Nil-Nil Theorem). The proof is a direct calculation on the Nil-categories and their \(K\)-theory. The first very interesting application of the result is an improvement of the Farrell-Jones Isomorphism Conjecture for \(K\)-theory. The original conjecture predicted that the \(K\)-theory of a group \(G\) can be calculated from the \(K\)-theory of its virtually cyclic subgroups. Using, their results the authors prove that we need to show that the \(K\)-theory of a group can be calculated from their finite-by-cyclic subgroups. As an application of their more refined version of the conjecture, the authors compute the \(K\)-theory of the modular group. The Nil-groups appear as obstructions to splitting problems in Topology. The authors define a map \(f: M \to X\) to be semisplit along \(N \subset M\), that separates \(M\), if the map induces an isomorphism on fundamental groups, and it is simple homotopic to a map \(h: (M, N) \to (X, Y)\) is, essentially, a homotopy equivalence of the one side of the splitting i.e., if \(M - N = M_1\cup M_2\), \(X - Y = X_1\cup X_2\), then \(h\) induces a homotopy equivalence between the pairs \((M_2, N)\) and \((X_2, Y)\). Their main result is that if \({\pi}(Y)\) has finite index in \({\pi}_(M_2)\), then \(f\) is semisplit. The results of this paper are the algebraic approach to similar results in [\textit{J. F. Davis, F. Quinn} and \textit{H. Reich}, J. Topol. 4, No. 3, 505--528 (2011; Zbl 1227.19004)].