A strictly decreasing invariant for resolution of singularities in dimension two (Q2179605)

The paper is in some sense a follow-up to the paper ``Canonical embedded and non-embedded resolution of singularities for excellent two-dimensional schemes'' by the first author et al. [``Canonical embedded and non-embedded resolution of singularities for excellent two-dimensional schemes'', Preprint, \url{arXiv:0905.2191}]. In that paper, it was proved that every reduced excellent Noetherian scheme \(X\) of dimension at most two admits a resolution of singularities via finitely many ``permissible'' blowups. Their resolution procedure (called the CJS algorithm) is ``canonical'' in a precise sense; in particular it commutes with localization and completion. The proof of termination is by contradiction, which is unusual in this business. The present paper introduces a new local invariant \(\iota(X, x)\) for schemes satisfying the above assumptions. In Theorem A, it is shown that this invariant decreases at each step of the CJS algorithm. This provides a new proof of termination of the CJS algorithm which is closer in spirit to the classical approach used in resolution theory.