Coordinate monotonicity and fixed suffixes survive polynomial limits
ProvedDerivativeChainPrivate.coordinate_chain_collision_limitarchive-supportconditional-resultsderivative-chains
For a fixed finite list of derivative orders, consider uniformly bounded lists of zeros of those iterated derivatives for coefficientwise-convergent polynomials with a common degree upper bound. A limiting list of derivative zeros retains convergent cost upper bounds, monotonicity of both coordinates and an eventually constant final suffix. Degree equality with the list length is not assumed, so this lemma alone does not assert a complete nonconstant-row tower or construct chains for arbitrary families.
Preamble
import Mathlib import Definitions.Def_DerivativeChainPrivate_foundations namespace DerivativeChainPrivate end DerivativeChainPrivate open DerivativeChainPrivate Filter Topology Polynomial open scoped BigOperators set_option autoImplicit false set_option backward.isDefEq.respectTransparency false
Formal statement
theorem DerivativeChainPrivate.coordinate_chain_collision_limit {n N : ℕ}
(p : ℕ → Polynomial ℂ) (q : Polynomial ℂ)
(hp : ∀ j, (p j).natDegree ≤ N) (hq : q.natDegree ≤ N)
(hc : ∀ k, Tendsto (fun j => (p j).coeff k) atTop (𝓝 (q.coeff k)))
(z : ℕ → Fin n → ℂ) (R : ℝ) (hR : 0 ≤ R)
(hzbound : ∀ j i, ‖z j i‖ ≤ R)
(hzroot : ∀ j i, (Polynomial.derivative^[n - (i.val + 1)] (p j)).eval (z j i) = 0)
(c : ℕ → ℝ) (C : ℝ) (hcLim : Tendsto c atTop (𝓝 C))
(hcost : ∀ j, chainCost (z j) ≤ c j)
(hre : ∀ j, Monotone (fun i => (z j i).re))
(him : ∀ j, Monotone (fun i => (z j i).im))
(k : ℕ) (last : Fin n) (htail : ∀ j i, k ≤ i.val → z j i = z j last) :
∃ w : Fin n → ℂ,
(∀ i, (Polynomial.derivative^[n - (i.val + 1)] q).eval (w i) = 0) ∧
chainCost w ≤ C ∧ Monotone (fun i => (w i).re) ∧
Monotone (fun i => (w i).im) ∧ (∀ i, k ≤ i.val → w i = w last) := by sorrySource