Bounded polynomial systems for auxiliary-function arguments
DefinitionTranscendenceTheory_BoundedBivariateSystemFix complex numbers , real constants , and a nonnegative integer . A bounded bivariate system consists of an matrix and a matrix over , with degree bounds , satisfying
Every entry of both matrices has -degree at most , -degree at most , and integer coefficients of absolute value at most . For every nonzero polynomial vector with those degree bounds, coefficient bound , and as a polynomial identity, some row of satisfies
The matrices encode vanishing equations and candidate derivative values in an auxiliary-function proof. The final property is an explicit obligation for a system constructor: the record does not prove the analytic interpolation or zero estimate. A separate coefficient-selection theorem constructs a bounded nonzero vector in the kernel.
Formalization Note. The outer polynomial variable is . Matrix sizes and degree bounds may depend on . The test matrix is chosen before the coefficient vector; its successful row may depend on that vector.
import Mathlib.Analysis.Complex.Polynomial.Basic
import Mathlib.Data.Matrix.Basic
open scoped Polynomial
namespace TranscendenceTheory
/-- The finite linear equations and test linear forms in an auxiliary-function argument.
The last field records the analytic nonvanishing and small-value estimates; constructing
such a system is a separate mathematical obligation. -/
structure BoundedBivariateSystem (θ ν : ℂ) (a c : ℝ) (N : ℕ) where
rows : ℕ
cols : ℕ
tests : ℕ
xDegree : ℕ
yDegree : ℕ
cols_pos : 0 < cols
dimension_gap : 8 * rows ≤ cols
xDegree_le : (xDegree : ℝ) ≤ a * N
yDegree_le : (yDegree : ℝ) ≤ a * N
size_le : (cols : ℝ) * (yDegree + 1) * (xDegree + 1) ≤ Real.exp (a * N)
equations : Matrix (Fin rows) (Fin cols) ℤ[X][X]
testForms : Matrix (Fin tests) (Fin cols) ℤ[X][X]
equations_yDegree : ∀ r i, (equations r i).natDegree ≤ yDegree
equations_xDegree : ∀ r i j, ((equations r i).coeff j).natDegree ≤ xDegree
equations_height : ∀ r i j k, ‖((equations r i).coeff j).coeff k‖ ≤ Real.exp (a * N)
testForms_yDegree : ∀ r i, (testForms r i).natDegree ≤ yDegree
testForms_xDegree : ∀ r i j, ((testForms r i).coeff j).natDegree ≤ xDegree
testForms_height : ∀ r i j k, ‖((testForms r i).coeff j).coeff k‖ ≤ Real.exp (a * N)
small_nonzero : ∀ p : Fin cols → ℤ[X][X], p ≠ 0 →
(∀ i, (p i).natDegree ≤ yDegree) →
(∀ i j, ((p i).coeff j).natDegree ≤ xDegree) →
(∀ i j k, ‖((p i).coeff j).coeff k‖ ≤ Real.exp (2 * a * N)) →
(∀ r, ∑ i, equations r i * p i = 0) →
∃ r : Fin tests,
(∑ i, testForms r i * p i).eval₂ (Polynomial.aeval θ).toRingHom ν ≠ 0 ∧
‖(∑ i, testForms r i * p i).eval₂ (Polynomial.aeval θ).toRingHom ν‖ ≤
Real.exp (-c * (N : ℝ) ^ 2 * Real.log N)
end TranscendenceTheory