Verdier-Riemann-Roch for Chern class and Milnor class. (Q700531)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Verdier-Riemann-Roch for Chern class and Milnor class. |
scientific article |
Statements
Verdier-Riemann-Roch for Chern class and Milnor class. (English)
0 references
22 October 2002
0 references
There are several approaches to construct Chern classes for singular varieties, like the Chern-Schwartz-MacPherson class and the Fulton-Johnson class, which generalize the classical Chern class of the tangent bundle in the nonsingular case. The difference of both invariants is measured by the so-called Milnor class, which carries information of the singularities (for more details, see the author's survey [in: Algebraic geometry: Hirzebruch 70. Proc. alg. geom. conf. in honor of F. Hirzebruch's 70th birthday, Warszawa, Poland 1998. Contemp. Math. 241, 349--369 (1999; Zbl 0966.14004)]). These classes are closely related to the Verdier-Riemann-Roch problem, which reflects the contravariant aspect of the two theories \(K_0\) and \(H_\mathbb{Q}\), while the covariant aspect is reflected by the singular Riemann-Roch of Baum, Fulton and MacPherson. In the paper under review, the author highlights the bivariant viewpoint of the relation between Milnor classes and the Riemann-Roch theorems. Let \(f: X\to Y\) be a l.c.i. morphism, \(c_*\) the morphism given by MacPherson's natural transformation, and \(T_f\) the relative tangent bundle of \(f\). The author discusses the existence of a constructible function \(\alpha_f\in F(X)\) such that, for all \(\beta \in F(Y)\), \(c(T_f)\cap f^*(c_*(\beta)) = c_*(\alpha_f\cdot f^*(\beta))\) holds. It is shown that such an \(\alpha_f\) doesn't exist in general; for instance, counterexamples are provided by special blow-ups. However, the author proves the existence in the case of trivial fiber bundles with l.c.i. fiber, and conjectures that the answer is also affirmative for any Zariski locally trivial fiber bundle with l.c.i. fiber. He outlines a proof for the latter result up to a technical problem.
0 references