Globalizing F-invariants (Q2421252)

In this nicely written paper, the authors extend the notion of the Hilbert-Kunz multiplicity and the \(F\)-signature to a non-local \(F\)-finite ring; both defined via limits which are shown to exist. As one would hope, these global versions of Hilbert-Kunz multiplicity and \(F\)-signature agree with the previously defined local versions for local \(F\)-finite local rings. They show that these new global versions satisfy the same properties that the original local versions satisfy such as: (1) \(R\) is regular if and only if \(e_{HK}(R)=s(R)=1\). (2) There is a positive \(\delta\) depending only on the dimension of \(R\) such that \(R\) is regular if and only if \(e_{HK}(R) \leq 1 +\delta\) if and only if \(s(R) \geq 1-\delta\). (3) If \(R \rightarrow T\) is faithfully flat then the following inequalities hold: \(e_{HK}(R) \leq e_{HK}(T)\) and \(s(R) \geq s(T)\). (4) \(R\) is strongly \(F\)-regular if and only if \(s(R)>0\). (5) Let \(d\) be the Krull dimension of \(R\) and \(e(R_P)\) be the Hilbert-Samuel multiplicity of \(R_P\) and set \(e=\text{max}\{e(R_P) \mid P \in \text{Spec}(R)\}\). If \(e_{HK}(R) \leq 1 +\text{max}\{ 1/d!, 1/e\}\), then \(R\) is strongly \(F\)-regular and Gorenstein. The authors include several examples and remarks which greatly add to the readability of the paper.