Every monic complex cubic admits a bounded actual forward law
ProvedCubicTurning.every_monic_cubic_uniform_forward_lawFor complex , set , using the total real division convention that division by zero is . Let mean or the same two inequalities with exchanged. Define when holds, and otherwise. All threshold inequalities are non-strict, and all complex moduli are Euclidean. For every polynomial that is monic and has natural degree exactly , there exist complex numbers such that . Put , , and . For a labelled triple and , put . A law consists of three labelled paths and three complex axes , indexed by , with , , , , , , , , and , . Its mean energy is . This is finite labelled averaging with weights ; repeated endpoint values retain separate labels and may have identical paths. For these same , there exists such a law with common start satisfying, additionally and explicitly, for every , where the derivative orders are respectively , and satisfying and . Both inequalities are non-strict. The root enumeration and law are existential and need not be unique; repeated roots, a triple root, zero variance, and zero displacements are included. The zero polynomial and lower-degree polynomials are excluded by the hypotheses. No assumption that roots are real, no squarefree condition, and no greedy requirement is made. The terminal distribution is the equal-weight distribution on the three root slots with multiplicity.
import Mathlib import Definitions.Def_CubicTurning_foundations import Definitions.Def_DerivativeChainPrivate_foundations set_option autoImplicit false set_option maxHeartbeats 1200000 open scoped BigOperators open ComplexConjugate
theorem CubicTurning.every_monic_cubic_uniform_forward_law (p : Polynomial ℂ)
(hp : p.Monic) (hd : p.natDegree=3) :
∃ a b c : ℂ, p=DerivativeChainPrivate.fromRoots (![a,b,c] : Fin 3 → ℂ) ∧
∃ L : AffineUniformForwardLaw a b c (centroid a b c),
(∀ i j : Fin 3, ((Polynomial.derivative^[3-(j.val+1)]) p).eval (L.paths i j)=0) ∧
L.meanEnergy ≤ (anisotropy a b c +
Real.sqrt (anisotropy a b c*(2*centeredVariance a b c-anisotropy a b c)))/3 ∧
L.meanEnergy ≤ (2/3)*centeredVariance a b c := by sorry