Quantitative \(K\)-theory for Banach algebras (Q1675913): Difference between revisions

Group C*-algebras and crossed products for group actions come with a canonical filtration by functions on the group supported on finite subsets. Quantitative \(K\)-theory is a variant of \(K\)-theory that takes such a filtration into account. This has been used to prove the Novikov conjecture for groups with finite asymptotic dimension and finite classifying space. Here quantitative \(K\)-theory is defined for certain Banach algebras. Applications are not treated. They appear in a separate article, which deals with the \(K\)-theory of Banach algebraic crossed products. Quantitative \(K\)-theory uses explicit norms on all matrix algebras. Hence the author restricts to Banach algebras of operators on subquotients of \(L^p\)-spaces for some \(p\), so as to have canonical norms on all matrix algebras. In addition, a filtration \(A=\bigcup A_r\) with vector subspaces \(A_r\) satisfying \(A_r A_s \subseteq A_{r+s}\) is fixed. Quantitative \(K_0\) considers matrices \(p\) with entries in \(A_r\) such that \(p^2-p\) has norm at most~\(\varepsilon\), and quantitative \(K_1\) considers pairs of matrices \(u,v\) with entries in \(A_r\) so that \(u v - 1\) and \(v u - 1\) has norm at most~\(\varepsilon\). These are considered up to homotopy. Letting \(r\) go to infinity and \(\varepsilon\) to \(0\), one recovers the usual K-theory of \(A\). The article defines the even and odd quantitative \(K\)-theory and proves its basic properties in great detail. This includes an exact sequence for an extension, but not Bott periodicity, and a Mayer-Vietoris sequence. The Mayer-Vietoris sequence works for two subalgebras with certain properties, not just for ideals.