Skip to content

Latest commit

 

History

History
47 lines (37 loc) · 2.7 KB

README.md

File metadata and controls

47 lines (37 loc) · 2.7 KB

Weights

Formalization in Lean4 of some results in Minimization of hypersurfaces by A.-S. Elsenhans and myself.

Contents

The paper introduces the notion of a weight on forms of degree $d$ in $n+1$ variables, which is simply an $(n+1)$-tuple of natural numbers. We index the entries from $0$ to $n$. See Weight.

We can then compare weights in the following way. We first define $J_{n,d}$ to be the Set of $(n+1)$-tuples of natural numbers whose sum is $d$ (Weight.testvecs n d). Then to a weight $w$, we associate the map $f_w \colon J_{n,d} \to \mathbb{Z}_{\ge 0}$ given by sending $a \in J_{n,d}$ to $$\max\Bigl\{0, \Bigl\lfloor\frac{d , \Sigma w}{n+1}\Bigr\rfloor - \langle a, w \rangle\Bigr\}$$ (where $\Sigma w$ denotes the sum of the entries of $w$). See Weight.f.

Then we say that a weight $w$ dominates another weight $w'$ if $f_w \le f_{w'}$ point-wise. In this file, we write w ≤d w' for this relation. ≤d is a pre-order on the set of weights, but not a (partial) order. For example, a weight $w$ and $w + k \mathbf{1}$ dominate each other for each natural number $k$. We can therefore restrict to weights whose minimal entry is zero.

We say that a weight $w$ that is (weakly) increasing and such that $w_0 = 0$ is normalized (Weight.normalized). We show that if a normalized weight $w$ dominates some weight $w'$, then it also dominates the increasing permutation of $w'$ (Weight.dom_of_dom_perm). So up to permutations, it suffices to consider only normalized weights.

We say that a set $S$ of normalized weights is complete if every normalized weight is dominated by an element of $S$ (Weight.complete_set). We say that a complete set of weights is minimal if it is minimal with respect to inclusion among complete sets (Weight.minimal_complete_set). This is equivalent to saying that when $w$ and $w'$ are in $S$ and $w$ dominates $w'$, then $w = w'$.

The first main result formalized here is that there is a unique minimal complete set of weights, which is given by the set Weight.M n d of all normalized weights that are minimal elements with respect to domination within the set of all normalized weights. This is Proposition 3.13 in the paper. See Weight.M_is_minimal and Weight.M_is_unique.

We show in addition that the entries of nonzero elements of M n d are coprime (Weight.gcd_eq_one_of_in_M) and that M n 1 consists of the single element $(0,1,\ldots,1)$ (Weight.w1_unique).

The second main result is a proof of Theorem 1.6 in the paper, which says that in the case $n = 2$, the weights in a minimal complete set of normalized weights have entries bounded by the degree $d$. See Weight.dom_by_max_le_d and Weight.theorem_1_6.