\(K_1\)-groups via binary complexes of fixed length (Q2284283): Difference between revisions

For \(\mathcal{N}\) an exact category, \textit{D. R. Grayson} [J. Am. Math. Soc. 25, No. 4, 1149--1167 (2012; Zbl 1276.19003)] defined \(K_1 (\mathcal{N})\) as a quotient of the Grothendieck group of binary acyclic complexes in \(\mathcal{N}\) and showed that it is naturally isomorphic to Quillen's \(K_1\). The authors refine this by using complexes of fixed length. For \(k \in \mathbb{N}\) they define \( K_1^k (\mathcal{N})\) by restricting to complexes supported on \([0,k]\), so that \(K_1 (\mathcal{N})\) is the colimit of these. The quotient \(L_1^k (\mathcal{N})\) is defined by imposing the binary ladder relation. The main technical ingredient of the paper is the proof that the natural homomorphism \(i_k : L_1^k (\mathcal{N})\rightarrow L_1^{k+1} (\mathcal{N})\) is an isomorphism for \(k \geq 2\). This is established using Grayson shortening, which is also exploited in [\textit{D. Kasprowski} and \textit{C. Winges}, Trans. Am. Math. Soc., Ser. B 7, 1--23 (2020; Zbl 07183274)]. The homomorphism \(i_k\) factorizes across the canonical surjection \(K_1^k (\mathcal{N}) \twoheadrightarrow L_1^k (\mathcal{N})\). This implies the main theorem of the paper, that the canonical map \[ L_1^k (\mathcal{N}) \rightarrow K_1 (\mathcal{N}) \] is an isomorphism for \(k \geq 2\). This allows a direct comparison with \textit{A. Nenashev}'s [J. Pure Appl. Algebra 131, No. 2, 195--212 (1998; Zbl 0923.19001)] \(K_1^N (\mathcal{N})\), which is defined using binary complexes of length two. Namely the above isomorphism for \(k \geq 2\) factorizes as \[ L_1^2 (\mathcal{N}) \twoheadrightarrow K_1^N (\mathcal{N}) \rightarrow K_1 (\mathcal{N}), \] so that both are isomorphisms. The authors conclude by explaining how Nenashev's relation follows from the binary ladder relation.