Algebraic analogue of the Atiyah completion theorem (Q2263822): Difference between revisions

Let \(G\) be a split reductive algebraic group and \(k\) be a base field. Let \(BG=\cup BG_m\) be the étale classifying space of \(G\), where \(BG_m=EG_m/G\) and the \(EG_m\) are \(k\)-smooth algebraic varieties with a free \(G\) action. Moreover these \(G_m\) are related by closed \(G\)-equivariant embeddings \(i_m:(EG)_m\to (EG)_{m+1}\). The motivic space \(EG\) is \(\mathbb{A}^1\)-contractible and has a free action. On the other hand, the Borel construction induces a \(K^G_0(k)\)-module map \(\phi_m:K^G_m(k)\to K_n(BG_m)\) which is compatible with the embeddings \(i_m\), hence there is a \(K^G_0(k)\)-module map \(\Phi_n:K^G_n(k)\to K_n(BG)\). Let \(I_G\) be the kernel of the augmentation map \(K^G_0(k)\to K_0(k)=\mathbb{Z}\). The main theorem of this paper is that the maps \[ \widehat{K^G_n(k))_{I_G}}\to \widehat{K_n(BG)}_{I_G} \leftarrow K_n(BG), \] are \(K^G_0(k)\)-module isomorphisms, where the left hand side is induced by \(\Phi_n\) and the right hand side is induce by completion. This is the analogue of the classical Atiyah-Segal completion theorem in topological \(K\)-theory.