Boundedness of \((\epsilon, n)\)-complements for surfaces (Q2020420)

The theory of complements plays a significant role in the development of the Minimal Model Program. In the article, the authors prove Shokurov's complements conjecture for surfaces, with a generalization to non-Fano type. For general dimensions, they show that the conjecture implies several other conjectures. In the end, some interesting open problems are proposed. The main result states as follows. Fix \(\epsilon\in \mathbb{R}_{\ge 0}\), \(p, M\in \mathbb{N}\) and a set \(\Gamma\subseteq [0, 1]\) satisfying the D.C.C. condition. Then there exists an \(n\in \mathbb{N}\) depending only on the given data and being divisible by \(p\), such that for any surface pair \((X/Z\owns z, B)\) with the properties: \begin{itemize} \item the coefficients of \(B\) are in \(\Gamma\), \item \((X/Z\owns z, B)\) is \((\epsilon, \mathbb{R})\)-complementary, and \item either \(\epsilon =0\) or the multiplicity of any fibre of any minimal elliptic fibration of the minimal resolution of \(X\) over \(Z\) is bounded from above by \(M\), \end{itemize} there exists an \((\epsilon, n)\)-complement \((X/Z\owns z, B^+)\) for \((X/Z\owns z, B)\). When \(\epsilon>0\), \(-(K_X+B)\) is big and nef over \(Z\), and \(\Gamma=\{\frac{m-1}{m}\:|\: m\in \mathbb{N}\}\cup \{1\}\), the existence of \((\delta, n)\)-complement for \((X/Z\owns z, B)\) for some \(0<\delta<\epsilon\) was obtained by \textit{C. Birkar} [``Boundedness of \(\varepsilon\)-log canonical complements on surfaces'', Preprint, \url{arXiv:math.AG/0409254}].