Poincaré invariants (Q880164): Difference between revisions

The authors construct an obstruction theory for relative Hilbert schemes, compute it explicitly for relative Hilbert schemes of divisors on smooth projective varieties, and study its relations with Seiberg-Witten theory in the special case of curves on an algebraic surface. In section one, using the theory developed by J. Li-G. Tian and K. Behrend-B. Fantechi, the authors prove the existence of a relative obstruction theory for the relative Hilbert scheme \(\text{ Hilb}^P_{V/S}\) over \(S\). Here \(V \to S\) is a flat projective morphism, and \(P\) is a fixed Hilbert polynomial. An essential ingredient is Flenner's adjoint functor for the category of complexes of coherent sheaves. Moreover, for the case of divisors on smooth projective varieties, they calculate Flenner's adjoint functor and determine the complex in the obstruction theory for the relative Hilbert schemes. The rest of the paper is devoted to the special case of curves on an algebraic surface. In section two, a second obstruction theory is found. For a surface \(V\) and a class \(m \in H^2(V; \mathbb{Z})\), the virtual fundamental classes \([[\text{ Hilb}_V^m]]\) and \(\{\text{ Hilb}_V^m \}\) from the first and second obstruction theories respectively are compared. Let \(k\) be the canonical class of \(V\). It is shown that \([[\text{ Hilb}_V^m]] = \{\text{ Hilb}_V^m \}\) if \(p_g(V) = 0\) and that \([[\text{ Hilb}_V^m]] = 0\) if \(p_g(V) > 0\) and \(\text{ Hilb}_V^m \times_{\text{ Pic}_V^m} \text{ Hilb}_V^{k-m}\) is empty. In addition, a Porteous type formula for the Hilbert scheme of curves on surfaces is proved. In section three, the authors define the Poincaré invariant \[ (P_V^+, P_V^-): H^2(V;\mathbb{Z}) \to \Lambda^*H^1(V;\mathbb{Z})\times \Lambda^*H^1(V;\mathbb{Z}). \] The behaviour of the Poincaré invariant under deformation and blowup of the surface is analyzed. Furthermore, a wall-crossing formula for the Poincaré invariant is given. In section four, the authors carry out explicitly computations for ruled surfaces, elliptic surfaces and logarithmic transformations. In section five, the Poincaré invariant and Seiberg-Witten invariant of the surface \(V\) are compared. The authors state a conjecture relating these two types of invariants, reduce it to the case when \(V\) is a minimal surface, verify it for every surface \(V\) with Kodaira dimension less than \(2\).