Noncommutative localisation in algebraic \(K\)-theory. II. (Q2370617)

In a previous joint paper with \textit{A. Ranicki} [``Noncommutative localisation in algebraic \(K\)-theory. I'', Geom. Topol. 8, 1385--1425 (2004; Zbl 1083.18007)], the author proved the following general localization theorem for the \(K\)-theory of a noncommutative ring \(A\): Let \(\sigma^{-1}A\) denote the Cohn localization with respect to a set \(\sigma = \{s_i:P_i \to Q_i\}\) of homomorphisms of finitely generated projective (left) \(A\)-modules. If the localization \( \sigma^{-1}A\) is stably flat over \(A\), i.e., the groups \(\text{Tor}^A_n(\sigma^{-1}A, \sigma^{-1}A)\) vanish for all \(n >0\), then the \((-1)\)-connected cover of the homotopy fiber of \(K(A) \to K(\sigma^{-1}A)\) can be naturally identified with the \(K\)-theory \(K(\mathcal R)\) of the Waldhausen subcategory \(\mathcal R\), generated by \(\sigma\), of the category \({\mathcal C}^{\text{perf}}(A)\) of perfect complexes of \(A\)-modules. In the current paper the author shows that this result yields generalizations of various existing theorems on localization in \(K\)-theory. In [``Localization for the \(K\)-theory of noncommutative rings'', Contemp. Math. 126, 219--230 (1992; Zbl 0780.19007)], \textit{C. Weibel} and \textit{D. Yao} considered the special case that all elements of \(\sigma\) are maps \(s_i :A \to A\), and that \(\sigma^{-1}A\) is flat. The author gives examples where the localization \(\sigma^{-1}A\) is stably flat, but not flat, so that even in this special case the result of Neeman-Ranicki needs weaker assumptions. In the case that all maps in \(\sigma\) are monomorphisms the author defines an exact category \(\mathcal E\), whose objects are certain finitely presented \(A\)-modules of projective dimension \(\leq 1\), which can be viewed as \((A,\sigma)\)-torsion modules, and he proves that \(K(\mathcal R) \cong K(\mathcal E)\). This generalizes the classical localization theorem due to Quillen [cf. \textit{D. Grayson}, ``Higher algebraic K-theory. II'', in: Algebr. K-Theory, Proc. Conf. Evanston 1976, Lect. Notes Math. 551, 217--240 (1976; Zbl 0362.18015)]. If \(A \to \sigma^{-1}A\) is a stably flat localization, then the main result of Neeman-Ranicki mentioned above yields a long exact sequence \[ \cdots \to K_2(\sigma^{-1}A)\to K_1(\mathcal R) \to K_1(A) \to K_1(\sigma^{-1}A)\to K_0(\mathcal R) \to K_0(A) \to K_0(\sigma^{-1}A). \] The author shows that without any assumption the last six-term part of the sequence \[ K_1(\mathcal R) \to K_1(A) \to K_1(\sigma^{-1}A)\to K_0(\mathcal R) \to K_0(A) \to K_0(\sigma^{-1}A) \] is always exact. This generalizes a result of \textit{A. H. Schofield}'s [``Representation of rings over skew fields'', Lond. Math. Soc. Lect. Note Ser. 92 (1985; Zbl 0571.16001)].