Documentation

TauCeti.NumberTheory.Padics.Distribution.AnalyticStage

Analytic stages on ℤ_p and their evaluation #

For h : ℕ, the residue classes a ∈ ZMod (p ^ h) cut ℤ_p into the discs a + pʰℤ_p, where a is represented by its natural representative a.val < pʰ. On the disc of a the coordinate w = (x - a) / pʰ runs through ℤ_p, and a function analytic on each of these discs is a family of power series ∑ₙ c(a, n) wⁿ whose coefficients tend to zero. With coefficients in a complete ultrametric normed ℤ_p-module E, these coefficient families make up the analytic stage

LA_h(E) = C₀(ZMod (pʰ) × ℕ, E),

normed by the largest coefficient (the maximum of the Gauss norms on the discs). Its elements are coefficient vectors, kept distinct from the functions they define: evaluation need not be injective when E has p-torsion. If multiplication by p on E is injective, evaluation is injective: values at powers of p in each normalized disc coordinate determine all coefficients by cancellation and a vanishing tail estimate.

Evaluation TauCeti.AnalyticStage.eval sends a coefficient vector to the continuous function x ↦ ∑ₙ wⁿ • c(a, n). It is linear over every scalar ring acting on E compatibly with ℤ_p, and does not increase norms.

Passing from level h to level h + 1 subdivides each disc a + pʰℤ_p into the p discs of the classes a' = a + pʰ b, 0 ≤ b < p, in ZMod (p ^ (h + 1)). On the smaller disc w = b + p v, so re-expanding in v gives the coefficients

d(a', j) = ∑_{k ≥ j} (k choose j) b^(k - j) pʲ • c(a, k).

The transition TauCeti.AnalyticStage.transition is this re-expansion. It does not increase norms and it is compatible with evaluation. Composing these adjacent transitions gives transitionLE S hhk : LA_h(E) →L[S] LA_k(E) for every h ≤ k. These maps form a directed system, do not increase norms, and preserve evaluation. No injectivity assumption is needed. When multiplication by p on E is injective, every transition is injective: reading the coefficients on the subdisc with the same centre recovers p^j times the original coefficient.

The disc decomposition itself is the homeomorphism PadicInt.discHomeomorph of TauCeti.NumberTheory.Padics.Disc, which sends (a, w) to a.val + pʰ w.

Main definitions #

Main results #

References #

Analytic stages #

@[reducible, inline]
abbrev TauCeti.AnalyticStage (p h : ℕ) (E : Type u_1) [NormedAddCommGroup E] :
Type u_1

The analytic stage LA_h(E) = C₀(ZMod (pʰ) × ℕ, E) of level h: the coefficient vectors c(a, n) of families of power series ∑ₙ c(a, n) wⁿ, one on each disc a + pʰℤ_p, in the coordinate w = (x - a.val) / pʰ. Its norm is the largest coefficient norm.

Instances For

    Evaluation of the analytic stage LA_h(E): the coefficient vector c defines the continuous function equal to ∑ₙ wⁿ • c(a, n) on the disc a + pʰℤ_p, where w = (x - a.val) / pʰ. It is linear over every scalar ring S acting on E compatibly with ℤ_p.

    Instances For
      theorem TauCeti.AnalyticStage.hasSum_eval_apply {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (x : ℤ_[p]) :
      HasSum (fun (n : ℕ) => ((PadicInt.discHomeomorph p h).symm x).2 ^ n • c ((PadicInt.toZModPow h) x, n)) (((eval S) c) x)

      The value of the evaluation at x is the power series of the disc of x at the disc coordinate of x.

      theorem TauCeti.AnalyticStage.eval_apply {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (x : ℤ_[p]) :
      ((eval S) c) x = ∑' (n : ℕ), ((PadicInt.discHomeomorph p h).symm x).2 ^ n • c ((PadicInt.toZModPow h) x, n)

      The evaluation of c at x is the sum of the power series of the disc of x at the disc coordinate of x.

      theorem TauCeti.AnalyticStage.hasSum_eval_natCast_val_add_mul {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (a : ZMod (p ^ h)) (w : ℤ_[p]) :
      HasSum (fun (n : ℕ) => w ^ n • c (a, n)) (((eval S) c) (↑a.val + ↑p ^ h * w))

      On the disc a + pʰℤ_p, the evaluation of c is the power series ∑ₙ wⁿ • c(a, n) in the disc coordinate w.

      theorem TauCeti.AnalyticStage.eval_natCast_val_add_mul {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (a : ZMod (p ^ h)) (w : ℤ_[p]) :
      ((eval S) c) (↑a.val + ↑p ^ h * w) = ∑' (n : ℕ), w ^ n • c (a, n)

      On the disc of a, the evaluation of c at a.val + pʰ w is ∑' n, wⁿ • c(a, n).

      Evaluation does not increase norms.

      Evaluation has operator norm at most 1.

      Analytic-stage evaluation is injective if multiplication by p on the coefficient module is injective. The coefficient module need not carry a field structure or an isometric scalar action.

      Raising the level #

      The transition LA_h(E) → LA_{h+1}(E). The disc of a ∈ ZMod (pʰ) is the union of the discs of the classes a' = a + pʰ b, 0 ≤ b < p, in ZMod (p ^ (h + 1)), where b = ⌊a'.val / pʰ⌋. On the disc of a' the coordinates are related by w = b + p v, and the transition re-expands ∑ₖ c(a, k) wᵏ in powers of v:

      d(a', j) = ∑ₖ (k choose j) b^(k - j) pʲ • c(a, k).

      Instances For
        theorem TauCeti.AnalyticStage.transition_apply {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (a' : ZMod (p ^ (h + 1))) (j : ℕ) :
        ((transition S) c) (a', j) = ∑' (k : ℕ), (↑(k.choose j) * ↑(a'.val / p ^ h) ^ (k - j) * ↑p ^ j) • c (↑a'.val, k)

        The re-expansion formula for the transition.

        @[simp]
        theorem TauCeti.AnalyticStage.transition_apply_cast {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (a : ZMod (p ^ h)) (j : ℕ) :
        ((transition S) c) (a.cast, j) = ↑p ^ j • c (a, j)

        On the subdisc with the same natural centre, re-expansion multiplies the degree-j coefficient by p^j.

        Raising the level is injective if multiplication by p on the coefficient module is injective.

        theorem TauCeti.AnalyticStage.norm_transition_apply_le {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] (c : AnalyticStage p h E) (a' : ZMod (p ^ (h + 1))) (j : ℕ) :
        ‖((transition S) c) (a', j)‖ ≤ ‖↑p‖ ^ j * ‖c‖

        The coefficient of vʲ after the transition has norm at most ‖p‖ʲ ‖c‖.

        The transition does not increase norms.

        @[simp]

        The transition is compatible with evaluation: a coefficient vector and its re-expansion on the smaller discs define the same function.

        @[simp]

        Evaluation after the transition is evaluation, as an identity of continuous linear maps.

        noncomputable def TauCeti.AnalyticStage.transitionLE {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k : ℕ} (hhk : h ≤ k) :

        Re-expansion from level h to level k, obtained by composing adjacent transitions.

        Instances For
          @[simp]

          Re-expansion from a level to itself is the identity.

          Re-expansion to the next target level appends one adjacent transition.

          @[simp]

          Re-expansion to the adjacent level is the original transition.

          @[simp]

          Appending an adjacent transition is re-expansion to the next target level.

          @[simp]
          theorem TauCeti.AnalyticStage.transition_transitionLE {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k : ℕ} (hhk : h ≤ k) (c : AnalyticStage p h E) :
          (transition S) ((transitionLE S hhk) c) = (transitionLE S ⋯) c

          Applying an adjacent transition after re-expansion raises the target level by one.

          @[simp]
          theorem TauCeti.AnalyticStage.transitionLE_comp {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k l : ℕ} (hhk : h ≤ k) (hkl : k ≤ l) :

          Re-expansion through an intermediate level equals the transition to the final level.

          @[simp]
          theorem TauCeti.AnalyticStage.transitionLE_transitionLE {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k l : ℕ} (hhk : h ≤ k) (hkl : k ≤ l) (c : AnalyticStage p h E) :
          (transitionLE S hkl) ((transitionLE S hhk) c) = (transitionLE S ⋯) c

          Applying two successive transitions is the transition across the combined levels.

          @[simp]

          Re-expansion after an adjacent transition starts one level earlier.

          @[simp]
          theorem TauCeti.AnalyticStage.transitionLE_transition {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k : ℕ} (hkl : h + 1 ≤ k) (c : AnalyticStage p h E) :
          (transitionLE S hkl) ((transition S) c) = (transitionLE S ⋯) c

          Applying re-expansion after an adjacent transition starts one level earlier.

          @[simp]

          Two adjacent transitions compose to re-expansion across two levels.

          @[simp]

          Two adjacent transitions re-expand a coefficient vector across two levels.

          instance TauCeti.AnalyticStage.instDirectedSystemTransitionLE {p : ℕ} [hp : Fact (Nat.Prime p)] {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] :
          DirectedSystem (fun (h : ℕ) => AnalyticStage p h E) fun {x x_1 : ℕ} (hhk : x ≤ x_1) => ⇑(transitionLE S hhk)

          Analytic-stage transitions form a directed system of coefficient vectors.

          Re-expansion between any two analytic levels does not increase the Gauss norm.

          Re-expansion between any two analytic levels has operator norm at most 1.

          @[simp]
          theorem TauCeti.AnalyticStage.eval_transitionLE {p : ℕ} [hp : Fact (Nat.Prime p)] {h : ℕ} {E : Type u_1} [NormedAddCommGroup E] [Module ℤ_[p] E] [IsBoundedSMul ℤ_[p] E] [IsUltrametricDist E] [CompleteSpace E] (S : Type u_2) [Semiring S] [Module S E] [ContinuousConstSMul S E] [SMulCommClass ℤ_[p] S E] {k : ℕ} (hhk : h ≤ k) (c : AnalyticStage p h E) :
          (eval S) ((transitionLE S hhk) c) = (eval S) c

          Re-expansion between any two analytic levels preserves the evaluated function.

          @[simp]

          Evaluation after re-expansion between two levels is evaluation, as an identity of continuous linear maps.

          Re-expansion between any two levels is injective if multiplication by p on the coefficient module is injective.