ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let W(b‾;a‾)W(\underline{b};\underline{a}) denote the locus of good determinantal schemes in Pn\mathbb{P}^{n} of codimension cc defined by the maximal minors of a homogeneous matrix with degree data b‾\underline{b} and a‾\underline{a}. Let

λc:=∑i,j(ai−bj+nn)+∑i,j(bj−ai+nn)−∑i,j(ai−aj+nn)−∑i,j(bi−bj+nn)+1.\lambda_c:= \sum_{i,j} \binom{a_i-b_j+n}{n} + \sum_{i,j} \binom{b_j-a_i+n}{n} - \sum _{i,j} \binom{a_i-a_j+n}{n}- \sum _{i,j} \binom{b_i-b_j+n}{n} + 1.

Given integers a0≤a1≤⋯≤at+c−2a_0\le a_1\le \cdots \le a_{t+c-2} and b1≤⋯≤btb_1\le\cdots\le b_t, set ℓi:=∑j=0t+i−2aj−∑k=1tbk\ell_i:=\sum_{j=0}^{t+i-2}a_j-\sum_{k=1}^tb_k and hi−3:=2at+i−2−ℓi+nh_{i-3}:=2a_{t+i-2}-\ell_i+n for i=3,4,…,ci=3,4,\ldots,c. Assume ai−min⁡([c/2]+1,t)≥bia_{i-\min([c/2]+1,t)}\ge b_i for min⁡([c/2]+1,t)≤i≤t\min([c/2]+1,t)\le i\le t.

The dimension conjecture for good determinantal schemes. Under these assumptions,

dim⁡W(b‾;a‾)=λc+K3+K4+⋯+Kc,\dim W(\underline{b};\underline{a})=\lambda_c+K_3+K_4+\cdots+K_c,

where K3=(h0n)K_3=\binom{h_0}{n},

K4=∑j=0t+1(h1+ajn)−∑i=1t(h1+bin),K_4=\sum_{j=0}^{t+1}\binom{h_1+a_j}{n}-\sum_{i=1}^{t}\binom{h_1+b_i}{n},

and, for 0≤i≤c−30\le i\le c-3,

Ki+3=∑r+s=ir,s≥0∑0≤i1<⋯<ir≤t+i1≤j1≤⋯≤js≤t(−1)i−r(hi+ai1+⋯+air+bj1+⋯+bjsn).K_{i+3}=\sum_{r+s=i\atop r,s\ge0}\sum_{0\le i_1<\cdots<i_r\le t+i\atop 1\le j_1\le\cdots\le j_s\le t}(-1)^{i-r}\binom{h_i+a_{i_1}+\cdots+a_{i_r}+b_{j_1}+\cdots+b_{j_s}}{n}.

In particular, when all entries of the matrix have the same degree, this specializes to dim⁡W(0‾;d‾)=t(t+c−1)(d+nn)−t2−(t+c−1)2+1\dim W(\underline{0};\underline{d})=t(t+c-1)\binom{d+n}{n}-t^2-(t+c-1)^2+1. The formula predicts the dimension of the determinantal locus and is known in some cases, including the ranges cited in the surrounding discussion, but is not established in general.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.