Operational \(K\)-theory (Q273499): Difference between revisions

\noindent For a quasi-projective scheme \(X\) let \(K^{0}(X)\) be the Grothendieck group of vector bundles on \(X\) and let \(K_{0}(X)\) be the Grothendieck group of coherent sheaves on \(X\). These groups are isomorphic whenever \(X\) is smooth. \textit{W. Fulton} and \textit{R. MacPherson} [Mem. Am. Math. Soc. 243, 165 p. (1981; Zbl 0467.55005)] introduced bivariant \(K\)-theory \(K(X\rightarrow Y)\) as a group of perfect complexes on \(X\). These groups specialize to the covariant functor \(K_{0}(X)=K^{0}(X\rightarrow \mathrm{pt})\) and the contravariant functor \(K^{0}(X)=K^{0}(X\mathop{\rightarrow}\limits^{\mathrm{id}}X)\). These groups allows one to unify and simplify the proofs of Riemann-Roch theorems. However, it does not have good functorial properties. For example there are no obvious pullbacks of an \(f\)-perfect complex via an arbitrary fiber square. In the paper the authors introduce and study the operational bivariant theory associated to the covariant theory of Grothendieck groups of coherent sheaves. This construction is inspired by the operational Chow theory and it has many analogous geometric properties of it although the proofs are different. It agrees with Grothendieck groups of vector bundles on smooth varieties, admits a natural map from the Grothendieck group of perfect complexes on general varieties, satisfies descent for Chow envelopes, and is A1-homotopy invariant. The authors show that the operational \(K\)-theory of a complete linear variety is dual to the Grothendieck group of coherent sheaves. They also characterize the equivariant operational K-theory of an arbitrary toric variety. The paper is nicely written and the exposition is clear.