Initial ideals of weighted forms and the genus of locally Cohen-Macaulay curves
Alessio Sammartano, Enrico Schlesinger
TL;DR
The paper resolves the sharpness of the upper bound for the maximum genus of locally Cohen-Macaulay curves in P^3 by translating the geometric question into the weighted-algebraic problem of initial ideals in a nonstandard graded ring with weights wt(x)=1,wt(y)=2,wt(z)=3. It proves that for a general wt-homogeneous form f of weight 3m, the multiplication map f·:R[−3m]→M is an isomorphism, which is equivalent to the initial ideal of (x,y)^{3m−2}+(f) being (x,y,z)^{3m−2}. The core method constructs a canonical bijection φ between monomial bases, decomposing into rectangular and triangular regions, and proving divisibility and, in favorable weights, uniqueness; for other weights, a refined Non Cancellation Lemma still yields the required isomorphism. The results confirm Beorchia–Lella–Sammartano’s conjecture in the d=s and d≥2s−1 cases, opening a path to the maximum genus predictions and linking algebraic combinatorics of weighted initial ideals to geometric genus bounds. These findings bridge weighted Gröbner-type questions with classical curve-genus problems, providing a rigorous combinatorial route to sharp genus bounds and offering explicit obstructions and remedies via special blocks and non-cancellation techniques.
Abstract
Let C be a locally Cohen-Macaulay curve in complex projective 3-space. The maximum genus problem predicts the largest possible arithmetic genus g(d,s) that C can achieve assuming that it has degree d and does not lie on surfaces of degree less than s. In this paper, we prove that this prediction is correct when d=s or d is at least 2s-1. We obtain this result by proving another conjecture, by Beorchia, Lella, and the second author, about initial ideals associated to certain homogeneous forms in a non-standard graded polynomial ring.
