Documentation

LeanPool.NandakumarRamanaRao.NRR.PrimePolyhedron.FoxNeuwirth.EquivariantPrismGenericPerturbation

Simultaneous generic perturbation of the equivariant refined prism #

The orbit-parameter construction already enforces shared-face compatibility and prime equivariance. The polynomial layer records all local facet determinants and codimension-two deviation minors, and all of those polynomials are nonzero. This module applies the finite multivariate perturbation mechanism at the assignment obtained by sampling a zero-free equivariant homotopy.

There is one quantitative subtlety. The parameter t in FiniteMultivariateGenericPerturbation.exists_small_positive_generic is small, but its target assignment is not a priori bounded, so that statement alone does not imply that the resulting assignment is close to the base assignment. We therefore use it once to obtain a generic target, then apply the finite univariate root-avoidance theorem on the line from the base assignment to that generic target. The second parameter is chosen using the finite L¹ size of the direction. The result is genuinely coordinatewise close to the homotopy assignment.

The final theorem assumes a positive coordinate form of the zero-free norm margin for the affine interpolation of the unperturbed homotopy samples. This is the exact quantitative hypothesis that must later be supplied by sufficiently fine spatial and staircase subdivision. The perturbation retains half of that margin, while polynomial nonvanishing gives facet regularity and codimension-two avoidance on every refined prism simplex.

A genuinely small generic multivariate perturbation #

Pointwise coordinate closeness of two finite parameter assignments.

Equations
Instances For
    theorem NRR.FoxNeuwirthOrderComplex.EquivariantPrismGenericPerturbation.exists_generic_assignment_close {J : Type u_1} {I : Type u_2} [Finite J] [Finite I] (P : I → MvPolynomial J ℝ) (hP : ∀ (i : I), P i ≠ 0) (a : J → ℝ) {eps : ℝ} (heps : 0 < eps) :
    ∃ (a' : J → ℝ), AssignmentClose a' a eps ∧ ∀ (i : I), (MvPolynomial.eval a') (P i) ≠ 0

    A finite family of nonzero multivariate polynomials admits a simultaneously nonvanishing assignment which is genuinely pointwise close to any prescribed base assignment.

    The first application of exists_small_positive_generic supplies a generic assignment. A second, univariate root-avoidance step moves an arbitrarily small positive distance from the base assignment toward that generic point.

    From polynomial nonvanishing to local general position #

    All ordered codimension-two deviation minors of one reconstructed local vertex map are nonsingular.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      Nonvanishing of the minor part of the combined polynomial family gives nonsingularity of every ordered codimension-two deviation matrix on one local refined prism simplex.

      The index of j after omitting the distinct index i.

      Equations
      Instances For
        theorem NRR.FoxNeuwirthOrderComplex.EquivariantPrismGenericPerturbation.sum_codimTwoVertex {p : ℕ} (hp : Nat.Prime p) (F : Fin (p + 1) → ℝ) (i j : Fin (p + 1)) (hij : i ≠ j) (hi : F i = 0) (hj : F j = 0) :
        ∑ k : Fin (p + 1), F k = ∑ c : Fin (p - 1), F (EquivariantPrismGenericityPolynomials.codimTwoVertex hp (i, omittedIndex i j hij) c)

        The ordered codimension-two vertex enumeration exhausts all vertices except the two omitted ones. This sum identity is the bridge from the minor determinant to barycentric avoidance.

        Nonsingularity of every ordered codimension-two deviation matrix rules out a deviation-zero point on a codimension-two face.

        Quantitative retention of the zero-free affine margin #

        A coordinate witness for a lower bound on the sup norm of every affine value. For the finite function space Fin p → Real, this is the concrete form needed to retain a zero-free norm margin under coordinatewise perturbation.

        Equations
        • One or more equations did not get rendered due to their size.
        Instances For

          A perturbation by less than eps in every parameter coordinate retains the affine coordinate norm margin, reduced from m to m - eps.

          A positive retained coordinate norm margin implies origin avoidance on every local affine simplex.

          The prism perturbation theorem #

          Output of the compatible equivariant generic perturbation. Shared-face compatibility and prime equivariance are inherited automatically from the orbit parameter space.

          Instances For

            Apply finite multivariate generic perturbation at the homotopy assignment. If the unperturbed piecewise-affine interpolation has positive coordinate norm margin m, the resulting compatible equivariant assignment has all facet determinants and codimension-two minors nonzero and retains margin m/2; hence every local prism simplex satisfies the full affine general-position interface.