Evaluation of is an inner product with a row
ProvedNearEnemy.eval_innerPolyevaluation-mapsinner-product-spacesnear-enemypolynomial-method
Let be a coefficient matrix (two rows indexed by ), and let be a vector in EuclideanSpace ℝ ι. Then evaluating the polynomial innerPoly k v at recovers the inner product of the -th row of with :
This evaluation identity is the semantic core of the polynomial encoding: the formal polynomial innerPoly faithfully represents the linear functional determined by a projection row. It lets the nonvanishing results proved for polynomials transfer to geometric statements about inner products with projection rows.
Preamble
import Mathlib
import Definitions.Def_NearEnemyDefs
universe u_1
open scoped RealInnerProductSpace
open scoped Classical
open MvPolynomial
variable {V : Type*} [NormedAddCommGroup V] [InnerProductSpace ℝ V]
variable {ι : Type*} [Fintype ι]
open NearEnemy
Formal statement
theorem NearEnemy.eval_innerPoly (f : Fin 2 × ι → ℝ) (k : Fin 2)
(v : EuclideanSpace ℝ ι) :
eval f (innerPoly k v) = ⟪rowOf f k, v⟫ := by sorry
Source
Prior art: Lund-Sheffer-de Zeeuw, Bisector energy and few distinct distances, SoCG 2015, LIPIcs vol. 34, 537-552, DOI 10.4230/LIPIcs.SOCG.2015.537, footnote 1 on p. 538, state that E(P) = 2n(n-1) when every pair of distinct points has a distinct perpendicular bisector, with the count of trivial quadruples that proves the floor (this footnote is not in arXiv:1411.6868v1); the asymptotic floor E(P) = Omega(n^2) is in their section 3.4. The generic planar projection that is injective, keeps general position and transports distances is Erdos-Furedi-Pach-Ruzsa, The grid revisited, Discrete Math. 111 (1993), proof of Theorem 3.1. Formalized in https://github.com/mysticflounder/lean-formalizations/blob/dd46c17a2a034d7bfa0df02e7f77834d35592864/lean/LeanFormalizations/Geometry/Euclidean/NearEnemyTheorem.lean#L986-L990