arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.04410math.AG
收藏

奇异曲线的微分Hodge结构I

Infinitesimal Variations of Hodge Structure for Singular Curves I

Xiamen University Malaysia(厦门大学马来西亚分校)

Mounir Nisse

AI总结:

本文研究了奇异曲线的微分Hodge结构,揭示了节点奇点与其它奇点在微分变化中的不同贡献,并在特定条件下给出了最大变化的判定标准。

摘要:

我们研究了具有奇点的代数曲线族所关联的微分Hodge结构。该分析不仅适用于节点情况,还适用于更一般的曲线,包括位于光滑项目曲面上的曲线以及具有更一般孤立奇点的族。利用变形理论和残差理论的方法,我们描述了微分周期映射如何分解为支持于奇点处的局部贡献,以及来自正则化的几何结构的全局约束。虽然节点奇点会产生非平凡的一阶贡献,其他奇点可能仅通过更高阶的局部数据贡献,或在微分层面不可见。因此,我们获得了以曲线的数值不变量为特征的最大微分变化的严格标准,特别是当节点数满足不等式δ≥g,其中g表示正则化的亏格。我们还把这些结果扩展到射影三维空间中非常一般的曲面上,证明在Picard秩为一的曲面上最大变化持续存在,但在存在更高ADE奇点的情况下,当亏格足够大时,最大变化失效。这些结果将经典微分Hodge理论中的最大性现象扩展到了更广泛的奇点和几何设置中。

英文摘要:

We study the infinitesimal variation of Hodge structure associated with families of reduced algebraic curves with singularities. The analysis applies to curves beyond the nodal case and is not restricted to plane curves, encompassing curves lying on smooth projective surfaces as well as families with more general isolated singularities. Using deformation-theoretic and residue-theoretic methods, we describe how the infinitesimal period map decomposes into local contributions supported at singular points, together with global constraints arising from the geometry of the normalization. While nodal singularities give rise to nontrivial rank-one contributions, other singularities may contribute only through higher-order local data or may be invisible at the infinitesimal level. As a consequence, we obtain sharp criteria for maximal infinitesimal variation in terms of numerical invariants of the curve, notably when the number of nodes satisfies the inequality $δ\ge g$, where $g$ denotes the genus of the normalization. We extend these results to curves on very general surfaces in projective three-space, showing that maximal variation persists on Picard-rank-one surfaces but fails for sufficiently large genus in the presence of higher ADE singularities. These results extend classical maximality phenomena in infinitesimal Hodge theory to a broader singular and geometric setting.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑