The weighted log canonical thresholds of toric plurisubharmonic functions (Q2256799)

The (ordinary) log canonical threshold is defined as \[ c_{\varphi}(z_{0}):=\sup\big\{c\in \mathbb R_{+}: e^{-2c\varphi}\text{ is locally integrable} \] \[ \text{ in some neighborhood of } z_{0}\big\}\in(0,\infty], \] for a plurisubharmonic function \(\varphi\) defined in a neighborhood of \(z_{0}\). The weighted log canonical threshold (which is studied in this paper) is the following generalization: \[ c_{\varphi,\mu}(z_{0}):=sup\big\{c\in \mathbb R_{+}: e^{-2c\varphi}\text{ is locally integrable} \] \[ \text{ in some neighborhood of } z_{0} \text{ with respect to the measure } d\mu\big\}\in(0,\infty], \] the measure \(\mu\) being assumed to be Radon. Toric plurisubharmonic functions are those that depend only on the absolute values of the coordinates of the argument. The main theorem says that negative toric plurisubharmonic functions \(\varphi\) on the polydisc can have their weighted log canonical threshold at zero computed explicitly as the inverse of the maximum of some of Kiselman's refined Lelong numbers (the measure has to obey some restrictions, which also determine which of the numbers one should consider). As a corollary, the weighted log canonical threshold is computed for measures of the type \(|f|^{2p}\) times the Lebesgue measure, where \(f\) is a germ of a holomorphic function such that it's multiplier ideal sheaf is generated by a given set of monomials.