Hilbert-Kunz multiplicity, McKay correspondence and good ideals in two-dimensional rational singularities (Q5937273)

The Hilbert-Kunz multiplicity \(e_{HK}(I)\) [see \textit{E. Kunz}, Am. J. Math. 91, 772-784 (1969; Zbl 0188.33702)], respectively \textit{P. Monsky} [Math. Ann. 263, 43-49 (1983; Zbl 0509.13023)], of an \(\mathfrak m\)-primary ideal \(I\) in a local ring \((A, \mathfrak m)\) of characteristic \(p > 0\) defined by the Frobenius morphism is known to be a rather mysterious invariant. The notion of a good ideal was introduced by \textit{S. Goto, S.-I. Iai} and \textit{K.-I. Watanabe} [see Trans. Am. Math. Soc. 353, No. 6, 2309-2346 (2001; Zbl 0966.13002)], as an ideal \(I\) in a local ring such that its associated graded ring \(\text{gr}_I(A)\) is a Gorenstein ring with \(a\)-invariant \(a(\text{gr}_I(A)) = 1-d.\) In the two-dimensional case it follows that \(I\) is good in \(A\) if and only if \(I\) is an integrally closed ideal represented on the minimal resolution \(f : X \to \text{Spec} A.\) The main result of the paper is a nice formula for the Hilbert-Kunz multiplicity for certain ideals \(I:\) Let \(A\) be a two dimensional F-finite rational Gorenstein local ring such that \(A\) is a pure subring of a regular ring \(B\) which is a finite \(A\)-module of rank \(N.\) Let \(I = H^0(X, \mathcal O_X(-Z))\) be a good ideal such that \(I\mathcal O_X = \mathcal O_X(-Z)\) with \(Z = \sum_{i=1}^r a_iE_i,\) where \(E_1, \ldots, E_r\) are irreducible exceptional curves on a minimal resolution of the singularity. Then \[ e_{HK}(I) - \text{Length}_A (A/I) = \left(\sum_{i=1}^r a_in_i\right) \Biggl/N, \] where the \(n_i\) is determined by the fundamental cycle \(Z_0 = \sum_{i=1}^r n_iE_i\) on \(X.\) Moreover \[ e_{HK}(I) - \text{Length}_A (A/I) = e_{HK}(\widetilde{I}) - \text{Length}_A (A/\widetilde{I}), \] where \(\widetilde{I}\) denotes the good closure of \(I,\) that is the minimal good ideal containing \(I.\) In the case \(A\) the completion of \(k[x,y]^G\), \(G\) a finite subgroup of \(SL(2,k),\) one can apply the result. So there is a description of \(e_{HK}(I)\) of any good ideal in terms of the dual graph of \(X.\) For the proof of their main results the authors use the McKay correspondance and the Riemann-Roch formula. The paper is completed by several lists of good ideals in two-dimensional rational singularities. By virtue of their main result there are calculations of \(e_{HK}(I)\) using its dual graph. So they provide a lot of new samples of the Hilbert-Kunz multiplicities of certain ideal.