Configuration spaces over Hilbert schemes and applications (Q2564770)

In this book the authors indicate a general procedure for obtaining multiple point formulas for morphisms \(f:V\to W\) of non-singular complex varieties, under quite general conditions. Intuitively a \(k\)-tuple point of \(f\) is a point \(x\) such that \(f^{-1}f(x)\) consists of at least \(k\) points, and the \(k\)-tuple locus is the subvariety of \(V\) consisting of all such points. A \(k\)-tuple point formula is a formula that expresses the class of the \(k\)-tuple locus in a suitable intersection theory for \(V\), in terms of polynomials in the Chern classes of the virtual normal bundle \(\nu(f) =f^\ast TW-TV\). For the formulas to be meaningful certain relations between the numbers \(\dim(V)\), \(\dim(W)\) and \(k\) must be satisfied. To obtain \(k\)-tuple formulas it is necessary to have a manageable parameter space for \(k\)-tuple points in \(V\). In this space one should be able to define a class of \(k\)-tuple points of \(f\), that, at least formally, represents the subvariety consisting of \(k\)-tuple points of \(V\) that are \(k\)-tuple points of the morphism \(f\). Unfortunately the variety consisting of \(k\)-tuple points of \(V\) that are \(k\)-tuple points of \(f\) often contain excess components that correspond to improper \(k\)-tuple points of \(f\). In order that the \(k\)-tuple formulas should have a geometric meaning it is necessary to give procedures for removing the improper ones. The authors give parameter spaces for the \(k\)-tuple points of \(V\) for \(k=2,3,4\), and define the class of \(k\)-tuple points of \(f\) in the Chow ring of these spaces. They also perform the necessary calculations in the Chow ring to obtain double- and triple-point formulas under quite general conditions on \(f\). We give some details of the construction of the parameter spaces used for double and triple points. To this end we denote by \(H^i(V)\) the Hilbert scheme of colength \(k\) points of a scheme \(V\). For the double points the authors use the two-sheeted cover \(\widehat{H^2(V)}\) of \(H^2(V)\). They use this space to obtain the well known double point formula [see, e.g., \textit{W. Fulton}, ``Intersection theory'' (1998; Zbl 0885.14002)]. The construction of the parameter space and the calculations to obtain the formula are performed in chapters 1 and 2. For the triple points the situation is much more complicated. The parameter space \(\widehat {H^3(V)}\) of triple points of \(V\) is the locus in \(V^3\times H^2(V)^3\times H^3(V)\) consisting of points of the form \((p_1, p_2, p_3, d_{12}, d_{23}, d_{13}, t)\) satisfying the relations: \[ \begin{cases} p_i\subset d_{ij} \subset t\\ p_j=\text{Res}(p_i,d_{ij})\\ p_k=\text{Res}(d_{ij},t) \quad \text{with} \quad \{i,j,k\} =\{1,2,3\}, \end{cases} \] where \(\text{Res}(\eta,\xi)\) denote the residual closed point of the \((k-1)\)-tuple \(\eta\) contained in the \(k\)-tuple \(\xi\). The authors follow a method of \textit{Z. Ran} [Acta Math. 155, 81-101 (1985; Zbl 0578.14046)] and \textit{T. Gaffney} [Math. Ann. 295, 269-289 (1993; Zbl 0841.14002)], and define a class that formally represents the triple points of \(f\). In chapter 3 they obtain the well known triple points formula of \textit{S. L. Kleiman} [Acta Math. 147, 13-49 (1993; Zbl 0479.14004)] and \textit{F. Ronga} [Compos. Math. 53, 211-223 (1984; Zbl 0563.57014)] under quite general conditions (see also \textit{W. Fulton}, loc. cit.). Many of the computations are tedious and are performed in chapter 4. The intersections of the locus of triple points corresponding to the class of triple points have excess intersections in the presence of \(S_2\)-singularities, and are difficult to handle. The authors indicate how this can be done, by treating the case \(\dim(V)=2\), \(\dim(W)=3\), when \(f\) has \(S_2\) singularities. A similar treatment of the triple points have been given by \textit{P. le Barz} [Duke Math. 5, 57, 925-946 (1988; Zbl 0687.14042) and Bull. Soc. Math. Fr. 112, 303-324 (1984; Zbl 0561.14021)]. The novel part of the book is the case of quadruple points. Here there is a real leap in difficulty. The second part of the book is devoted to this case. We shall not even indicate the construction of the parameter space. The reader who enjoys concrete, detailed geometry, and who likes to draw pictures representing ideals of the same colength, having various properties, will enjoy this part of the work.