Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement

Setting (p. 6). F\mathbb F is an arbitrary field and PG⁡(n,F)\operatorname{PG}(n,\mathbb F) the nn-dimensional projective geometry over it.

Theorem 5.1 (p. 7). Let tt be a positive integer and l1,…,lnl_1,\ldots,l_n concurrent lines through a point xx that span PG⁡(n,F)\operatorname{PG}(n,\mathbb F). Let SiS_i be a set of points of li∖{x}l_i\setminus\{x\} and DiD_i a proper nonempty subset of SiS_i. Suppose AA is a set of points such that every hyperplane ⟨s1,…,sn⟩\langle s_1,\ldots,s_n\rangle with (s1,…,sn)∈(S1×⋯×Sn)∖(D1×⋯×Dn)(s_1,\ldots,s_n)\in(S_1\times\cdots\times S_n)\setminus(D_1\times\cdots\times D_n) contains at least tt points of AA. If some hyperplane ⟨d1,…,dn⟩\langle d_1,\ldots,d_n\rangle with (d1,…,dn)∈D1×⋯×Dn(d_1,\ldots,d_n)\in D_1\times\cdots\times D_n contains no point of AA, then

∣A∣≥(t−1)max⁡j(∣Sj∣−∣Dj∣)+∑i=1n(∣Si∣−∣Di∣).|A|\ge(t-1)\max_j\bigl(|S_j|-|D_j|\bigr)+\sum_{i=1}^n\bigl(|S_i|-|D_i|\bigr).

The sets SiS_i are finite, as in the introduction (p. 1). The paper adds (pp. 7–8) that the proof gives the same bound for a multiset AA, that the hypothesis of a hyperplane missing AA cannot be dropped, and that for t=1t=1 the bound is attained by A=⋃i=1n(Si∖Di)A=\bigcup_{i=1}^n(S_i\setminus D_i). The introduction (pp. 1–2) illustrates the case n=2n=2, t=1t=1, where the bound is ∣S1∣+∣S2∣−2|S_1|+|S_2|-2 when D1D_1 and D2D_2 are single points.

Proof pointer

P. 7. A collineation sends a hyperplane through points of the DiD_i that misses AA to the hyperplane at infinity and the lines lil_i to the coordinate axes. The hyperplanes spanned by points of the SiS_i then become ∑itiXi=1\sum_i t_iX_i=1 with tit_i in parameter sets TiT_i, and f=∏a∈A(∑iaiXi−1)f=\prod_{a\in A}\bigl(\sum_i a_iX_i-1\bigr) has a zero of multiplicity tt outside the smaller parameter grid and is nonzero at the origin. Theorem 4.1 bounds deg⁡f\deg f. The print writes ∣A∣=deg⁡f|A|=\deg f; when x∈Ax\in A the factor of xx is constant, and the argument uses only deg⁡f≤∣A∣\deg f\le|A|.

Read depth

Claims checked: the statement and the remarks after it were read clause by clause against pp. 7–8 of the print, and the proof was followed.

Dependencies

Theorem 4.1.

Source. Simeon Ball and Oriol Serra, Punctured combinatorial Nullstellensätze, Combinatorica 29 (2009), 511–522, doi:10.1007/s00493-009-2509-z. Labels and page numbers are those of the corrected author manuscript dated 14 June 2011, the edition named on the source card.

Bears on

None recorded. The paper names no Erdős problem.