The maximum genus problem for locally Cohen-Macaulay space curves (Q1756604): Difference between revisions

One of the harder problems in the classification of space curves is the determination of the maximum genus of a smooth connected curve \(C \subset \mathbb P^3\) of degree \(d\) lying on no surface of degree \(< s\) (see the survey of \textit{R. Hartshorne} [Proc. Symp. Pure Math. 46, No. 1, 145--164 (1987; Zbl 0659.14020)] and the paper under review for the current status of this problem). The authors introduce an analogous problem of determining the maximal arithmetic genus \(P(d,s)\) of a locally Cohen-Macaulay curve \(C \subset \mathbb P^3\) of degree \(d\) lying on no surface of degree \(< s\). Since a curve of degree \(d < s\) lies on the cone over a general plane section, one need only consider curves satisfying \(d \geq s\). \textit{E. Schlesinger} has given an upper bound \(Q(d,s)\) on \(P(d,s)\) depending on whether \(s \leq d \leq 2s\) or \(d > 2s\) [Math. Nachr. 194, 197--203 (1998; Zbl 0929.14019)], extending the result of \textit{V. Beorchia} in characteristic zero [Math. Nachr. 184, 59--71 (1997; Zbl 0884.14012)]. For \(d>2s\) their bound also follows from bounds on the Rao function due to the reviewer and \textit{R. M. Miró Roig} [J. Pure Appl. Algebra 152, No. 1--3, 253--266 (2000; Zbl 0970.14016)]. The authors give evidence that \(P(d,s)=Q (d,s)\) for all \(d \geq s\). Observing that \(P(d-1,d-1)=Q (d-1,d-1)\) implies \(P(d,s) = Q (d,s)\) for \(s \geq 2d-1\), they focus on the case \(s \leq d \leq 2s\). Conjecture A asserts the existence of primitive multiplicity structures \(D\) supported on a line \(L\) with good cohomological properties that implies equality: recall that a \textit{primitive} multiplicity structure is one of embedding dimension two at each point, as defined by \textit{C. Bănică} and \textit{O. Forster} [Contemp. Math. 58, 47--64 (1986; Zbl 0605.14026)]. The main theorem gives sufficient conditions under which Conjecture A holds, reducing it to a Conjecture B about weighted polynomials in \(k[x,y,z]\) indexed by \(m \geq 2\). They prove Conjecture B for \(m=2\) and \(m=3\) by hand and confirm it for \(m \leq 40\) over a ground field of sufficiently large characteristic using the software \texttt{Macaulay2} of Grayson and Stillman.