Abstract Hilbert schemes. I (Q5949139)

In several previous papers [for example \textit{M. Artin} and \textit{J. J. Zhang}, Adv. Math. 109, 228-287 (1994; Zbl 0833.14002)], the authors have developed a theory of noncommutative algebraic geometry. Starting with a not necessarily commutative ring \(A\) (graded by the natural numbers \(\mathbb{N}\)) they do not define the scheme Proj(\(A\)) directly as a physical object, but instead define the category of ``coherent (respectively quasicoherent) sheaves on Proj(\(A\))'' as the quotient of the category of finitely generated (respectively arbitrary) graded \(A\)-modules by the subcategory of torsion modules (torsion means killed by some power of the graded maximal ideal of \(A\)). If \(A\) is commutative then under suitable hypotheses the resulting categories are equivalent to the categories of coherent or quasicoherent sheaves on the usual scheme Proj(\(A\)). The following is a somewhat simplified description of what the authors do in the present paper. They replace the categories of \(A\) modules involved in the construction of Proj by a \(k\)-linear Grothendieck category \({\mathcal C}\), where \(k\) is a commutative ring. Among other things the category \({\mathcal C}\) is abelian, and for objects \(V\) and \(W\) in \({\mathcal C}\), Hom\(_{\mathcal C}(V,W)\) is a \(k\)-module in a natural way. Typical examples are Proj(\(A\)) (regarded as an abelian category as described above), the category Mod(\(A\)) of right \(A\)-modules, and the category Gr(\(A\)) of graded right \(A\)-modules. If \(R\) is a Noetherian \(k\)-algebra and \(P\) is a Noetherian object in \({\mathcal C}\) then the Hilbert functor Quot\(_P\) associated to \(P\) assigns to \(R\) the set isomorphism classes of quotients of \(P\otimes_k R\) which are flat over \(R\). The authors' goal, following Grothendieck in his construction of the classical Hilbert scheme, is to show that (under suitable hypotheses) the Hilbert functor is representable. The paper is very long, much of it taken up by carefully stating the hypotheses necessary to carry out the authors' program and developing the necessary categorical machinery (tensor products, Hom, Tor, Ext, base change, completion, deformations). The authors' conclusion in section E3 is that under suitable hypotheses the Hilbert functor Quot\(_P\) is represented by a separated algebraic space locally of finite type over \(k\). More specifically (section E4) if \({\mathcal C}=\) Gr then the Hilbert functor of quotients with specified Hilbert function is represented by a projective scheme over \(k\), and (section E5) if \({\mathcal C}=\) Proj(\(A\)) the Hilbert functor is represented by a scheme which is a countable union of projective closed subschemes. The authors conjecture that the Hilbert functor of quotients with specified Hilbert function is represented by a projective scheme.