The coset bound `2p - 2` is attained: `x^{p-1} (xv)^{p-1}` is
ProvedHeisenberg125.Heis.productOneFree_cosetExtremalSeqaether-catalogalgebra
The coset bound 2p - 2 is attained: x^{p-1} (xv)^{p-1} is
product-one-free and lives in a single coset of the centre. In particular the
Erdős–Ginzburg–Ziv bound above cannot be improved.
theorem Heisenberg125.Heis.productOneFree_cosetExtremalSeq(hp : 0 < p) :
ProductOneFree (cosetExtremalSeq p) := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Algebra/Heisenberg125/CosetBound.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Algebra/Heisenberg125/CosetBound.lean
import Mathlib
import Definitions.Def_Algebra_Heisenberg125_Basic
import Definitions.Def_Algebra_Heisenberg125_CosetBound
import Definitions.Def_Algebra_Heisenberg125_LowerBound
/-
# A spread bound: at most `2p - 2` entries in any coset of the centre
The centre of `H_{p^3}` is `⟨v⟩`, and the cosets of the centre are the fibres of
the projection `(a,b,c) ↦ (a,b)` onto `(ZMod p)^2`.
**Theorem** (`Heis.length_le_of_const_image`). If `p` is odd and `C` is a
product-one-free sequence over `H_{p^3}` all of whose entries have the same
image `(α, β)` in `(ZMod p)^2`, then `|C| ≤ 2p - 2`.
The proof is a genuine cross-domain application of the **Erdős–Ginzburg–Ziv
theorem**: `2p - 1` entries in one coset contain `p` of them whose central
coordinates sum to zero; those `p` entries multiply (in any order) to
`(pα, pβ, Σc + (p choose 2) αβ) = 1`, using that `p` is odd so that
`p ∣ binom p 2`.
The bound is sharp: `v^{p-1}` and `x^{p-1}(xv)^{p-1}`-type sequences show that
`2p - 2` entries with equal image can occur inside a product-one-free sequence.
-/
open Heisenberg125
open Heis
variable {p : ℕ}
/-! ### Sharpness of the coset bound -/Formal statement
theorem Heisenberg125.Heis.productOneFree_cosetExtremalSeq(hp : 0 < p) :
ProductOneFree (cosetExtremalSeq p) := by sorrySource