Documentation

SHSLib.Gardiner.Ito.Increments

Wiener increments, quadratic variation and ∫W dW (Gardiner §4.2.2, §4.2.5) #

For a Wiener process W (Gardiner.Ito.IsWiener W P) and a partition t₀ = p 0 ≤ p 1 ≤ … ≤ p n = t of [t₀, t] (Gardiner.Ito.IsPartition, with mesh ≤ δ), write ΔWᵢ = W(p (i+1)) - W(p i) (Gardiner.Ito.dW) and Δtᵢ = p (i+1) - p i.

Erratum (4.2.18). The left-hand side is missing the square on Wᵢ - Wᵢ₋₁, and the middle line Σ_{i,j}[(tᵢ - tᵢ₋₁) - (t - t₀)][(tⱼ - tⱼ₋₁) - (t - t₀)] should be Σ_{i,j}(tᵢ - tᵢ₋₁)(tⱼ - tⱼ₋₁) - (t - t₀)², which vanishes. As printed, the middle line equals (n - 1)²(t - t₀)², so the printed right-hand side 2Σ Δtᵢ² + (n - 1)²(t - t₀)² is wrong (Gardiner.Ito.eq_4_2_18_printed_wrong: for n = 2 on [0, 1] the mean-square error is 1, the printed expression gives 2). The final value 2 Σ Δtᵢ² is correct.

Partitions #

structure Gardiner.Ito.IsPartition (p : ℕ → ℝ) (n : ℕ) (t₀ t δ : ℝ) :

p 0 = t₀ ≤ p 1 ≤ … ≤ p n = t is a partition of [t₀, t] with mesh at most δ (p is monotone on all of ℕ; its values beyond n are irrelevant).

  • start : p 0 = t₀
  • stop : p n = t
  • mono : Monotone p
  • mesh (i : ℕ) : i < n → p (i + 1) - p i ≤ δ
Instances For
    def Gardiner.Ito.unifPart (t₀ t : ℝ) (n i : ℕ) :

    The uniform partition tᵢ = t₀ + i(t - t₀)/n.

    Equations
    Instances For
      theorem Gardiner.Ito.isPartition_unifPart {t₀ t : ℝ} (h : t₀ ≤ t) {n : ℕ} (hn : 0 < n) :
      IsPartition (unifPart t₀ t n) n t₀ t ((t - t₀) / ↑n)
      theorem Gardiner.Ito.IsPartition.step_nonneg {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (i : ℕ) :
      0 ≤ p (i + 1) - p i
      theorem Gardiner.Ito.IsPartition.sum_step {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) :
      ∑ i ∈ Finset.range n, (p (i + 1) - p i) = t - t₀
      theorem Gardiner.Ito.IsPartition.le {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) :
      t₀ ≤ t
      theorem Gardiner.Ito.IsPartition.sum_sq_le {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) :
      ∑ i ∈ Finset.range n, (p (i + 1) - p i) ^ 2 ≤ δ * (t - t₀)

      Σᵢ Δtᵢ² ≤ δ (t - t₀).

      Increments #

      def Gardiner.Ito.dW {Ω : Type u_1} (W : ℝ → Ω → ℝ) (p : ℕ → ℝ) (i : ℕ) :
      Ω → ℝ

      The increment ΔWᵢ = W(p (i+1)) - W(p i) over the i-th cell of the partition p.

      Equations
      Instances For
        def Gardiner.Ito.qvSum {Ω : Type u_1} (W : ℝ → Ω → ℝ) (p : ℕ → ℝ) (n : ℕ) (ω : Ω) :

        The quadratic-variation sum Q = Σ_{i<n} ΔWᵢ².

        Equations
        Instances For
          def Gardiner.Ito.itoSum {Ω : Type u_1} (W : ℝ → Ω → ℝ) (p : ℕ → ℝ) (n : ℕ) (ω : Ω) :

          The left-point (Itô) sum S = Σ_{i<n} W(p i) ΔWᵢ of (4.2.12).

          Equations
          Instances For
            theorem Gardiner.Ito.psum_dW {Ω : Type u_1} (W : ℝ → Ω → ℝ) (p : ℕ → ℝ) (ω : Ω) (i : ℕ) :
            SDE.psum (fun (j : ℕ) => dW W p j ω) i = W (p i) ω - W (p 0) ω

            The partial sums of the increments are the displacements: Σ_{j<i} ΔWⱼ = W(p i) - W(p 0).

            theorem Gardiner.Ito.itoSum_eq {Ω : Type u_1} (W : ℝ → Ω → ℝ) (p : ℕ → ℝ) (n : ℕ) (ω : Ω) :
            itoSum W p n ω = 1 / 2 * (W (p n) ω ^ 2 - W (p 0) ω ^ 2) - 1 / 2 * qvSum W p n ω

            (4.2.12)–(4.2.13), for every path: S = ½[W(t)² - W(t₀)²] - ½ Σ ΔWᵢ². Derived from the algebraic identity Gardiner.SDE.ito_sum_identity of the Itô-free layer.

            theorem Gardiner.Ito.coe_step {p : ℕ → ℝ} (hp : Monotone p) (i : ℕ) :
            ↑(p (i + 1) - p i).toNNReal = p (i + 1) - p i
            theorem Gardiner.Ito.IsWiener.dW_measurable {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) (p : ℕ → ℝ) (i : ℕ) :
            Measurable (dW W p i)
            theorem Gardiner.Ito.IsWiener.dW_map {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :

            ΔWᵢ ~ N(0, Δtᵢ).

            theorem Gardiner.Ito.IsWiener.dW_iIndepFun {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) :

            The increments over a partition are independent.

            theorem Gardiner.Ito.IsWiener.dW_integral_comp {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) {f : ℝ → ℝ} (hf : Continuous f) :
            ∫ (ω : Ω), f (dW W p i ω) ∂P = ∫ (z : ℝ), f z ∂ProbabilityTheory.gaussianReal 0 (p (i + 1) - p i).toNNReal
            theorem Gardiner.Ito.IsWiener.dW_integrable_pow {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i k : ℕ) :
            MeasureTheory.Integrable (fun (ω : Ω) => dW W p i ω ^ k) P
            theorem Gardiner.Ito.IsWiener.dW_mean {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :
            ∫ (ω : Ω), dW W p i ω ∂P = 0

            ⟨ΔWᵢ⟩ = 0.

            theorem Gardiner.Ito.IsWiener.dW_sq_mean {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :
            ∫ (ω : Ω), dW W p i ω ^ 2 ∂P = p (i + 1) - p i

            (4.2.14), §4.2.5 i): ⟨ΔWᵢ²⟩ = Δtᵢ.

            theorem Gardiner.Ito.IsWiener.dW_four {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :
            ∫ (ω : Ω), dW W p i ω ^ 4 ∂P = 3 * (p (i + 1) - p i) ^ 2

            (4.2.17): ⟨ΔWᵢ⁴⟩ = 3Δtᵢ².

            theorem Gardiner.Ito.IsWiener.dW_sq_sub_sq {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :
            ∫ (ω : Ω), (dW W p i ω ^ 2 - (p (i + 1) - p i)) ^ 2 ∂P = 2 * (p (i + 1) - p i) ^ 2

            §4.2.5 ii): ⟨(ΔWᵢ² - Δtᵢ)²⟩ = 2Δtᵢ².

            theorem Gardiner.Ito.IsWiener.centred_memLp {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} (hp : Monotone p) (h0 : 0 ≤ p 0) (i : ℕ) :
            MeasureTheory.MemLp (fun (ω : Ω) => dW W p i ω ^ 2 - (p (i + 1) - p i)) 2 P

            The centred squared increments ΔWᵢ² - Δtᵢ.

            theorem Gardiner.Ito.IsWiener.qv_mean {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (h0 : 0 ≤ t₀) :
            ∫ (ω : Ω), qvSum W p n ω ∂P = t - t₀

            (4.2.14): ⟨Σ ΔWᵢ²⟩ = t - t₀.

            theorem Gardiner.Ito.IsWiener.qv_ms_error {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (h0 : 0 ≤ t₀) :
            ∫ (ω : Ω), (qvSum W p n ω - (t - t₀)) ^ 2 ∂P = 2 * ∑ i ∈ Finset.range n, (p (i + 1) - p i) ^ 2

            (4.2.15)–(4.2.18), corrected: the exact mean-square error of the quadratic-variation sum, ⟨(Σ ΔWᵢ² - (t - t₀))²⟩ = 2 Σ Δtᵢ².

            theorem Gardiner.Ito.IsWiener.qv_ms_le {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (h0 : 0 ≤ t₀) :
            ∫ (ω : Ω), (qvSum W p n ω - (t - t₀)) ^ 2 ∂P ≤ 2 * δ * (t - t₀)

            The mean-square error of the quadratic-variation sum is at most 2δ(t - t₀).

            theorem Gardiner.Ito.IsWiener.qv_msLim {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {t₀ t : ℝ} (h0 : 0 ≤ t₀) {p : ℕ → ℕ → ℝ} {n : ℕ → ℕ} {δ : ℕ → ℝ} (hp : ∀ (k : ℕ), IsPartition (p k) (n k) t₀ t (δ k)) (hδ : Filter.Tendsto δ Filter.atTop (nhds 0)) :
            Filter.Tendsto (fun (k : ℕ) => ∫ (ω : Ω), (qvSum W (p k) (n k) ω - (t - t₀)) ^ 2 ∂P) Filter.atTop (nhds 0)

            (4.2.19): ms-lim Σ ΔWᵢ² = t - t₀ along any sequence of partitions of [t₀, t] whose mesh tends to zero.

            theorem Gardiner.Ito.IsWiener.itoSum_ms_error {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (h0 : 0 ≤ t₀) :
            ∫ (ω : Ω), (itoSum W p n ω - 1 / 2 * (W t ω ^ 2 - W t₀ ω ^ 2 - (t - t₀))) ^ 2 ∂P = 1 / 2 * ∑ i ∈ Finset.range n, (p (i + 1) - p i) ^ 2

            The mean-square error of the left-point sum for ∫ W dW: ⟨(S - ½[(W(t)² - W(t₀)²) - (t - t₀)])²⟩ = ½ Σ Δtᵢ².

            theorem Gardiner.Ito.IsWiener.itoSum_msLim {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {t₀ t : ℝ} (h0 : 0 ≤ t₀) {p : ℕ → ℕ → ℝ} {n : ℕ → ℕ} {δ : ℕ → ℝ} (hp : ∀ (k : ℕ), IsPartition (p k) (n k) t₀ t (δ k)) (hδ : Filter.Tendsto δ Filter.atTop (nhds 0)) :
            Filter.Tendsto (fun (k : ℕ) => ∫ (ω : Ω), (itoSum W (p k) (n k) ω - 1 / 2 * (W t ω ^ 2 - W t₀ ω ^ 2 - (t - t₀))) ^ 2 ∂P) Filter.atTop (nhds 0)

            (4.2.20): ∫_{t₀}^t W dW = ms-lim Σ W(tᵢ₋₁)ΔWᵢ = ½[W(t)² - W(t₀)² - (t - t₀)] along any sequence of partitions whose mesh tends to zero.

            theorem Gardiner.Ito.IsWiener.ito_sum_of_qv_ms {Ω : Type u_1} [MeasurableSpace Ω] {W : ℝ → Ω → ℝ} {P : MeasureTheory.Measure Ω} (hW : IsWiener W P) {p : ℕ → ℝ} {n : ℕ} {t₀ t δ : ℝ} (hp : IsPartition p n t₀ t δ) (h0 : 0 ≤ t₀) :
            ∫ (ω : Ω), (∑ i ∈ Finset.range n, SDE.psum (fun (j : ℕ) => dW W p j ω) i * dW W p i ω - 1 / 2 * (SDE.psum (fun (j : ℕ) => dW W p j ω) n ^ 2 - (t - t₀))) ^ 2 ∂P = 1 / 2 * ∑ i ∈ Finset.range n, (p (i + 1) - p i) ^ 2

            Comparison with Gardiner.SDE.ito_sum_of_qv. That theorem needs the exact identity Σδᵢ² = t - t₀, which never holds for Wiener increments (it is a statement about a random variable). With real Wiener increments, ito_sum_of_qv holds up to an error whose mean square is exactly ½ Σ Δtᵢ²: for the increments δᵢ = ΔWᵢ(ω), Σ psum δ i · δ i = ½[(psum δ n)² - (t - t₀)] + Eₙ(ω) with ⟨Eₙ²⟩ = ½ Σ Δtᵢ² ≤ ½δ(t - t₀).

            Instances: the standard Wiener process on uniform partitions of [0, 1] #

            theorem Gardiner.Ito.sum_unif_sq (n : ℕ) (hn : 0 < n) :
            ∑ i ∈ Finset.range n, (unifPart 0 1 n (i + 1) - unifPart 0 1 n i) ^ 2 = 1 / ↑n
            theorem Gardiner.Ito.qv_unif_instance (n : ℕ) (hn : 0 < n) :
            ∫ (ω : StdWienerSpace), (qvSum stdWiener (unifPart 0 1 n) n ω - (1 - 0)) ^ 2 ∂stdWienerMeasure = 2 / ↑n

            Instance of (4.2.18)–(4.2.19): for the standard Wiener process and the uniform partition of [0, 1] into n cells, ⟨(Σ ΔWᵢ² - 1)²⟩ = 2/n, which tends to 0.

            theorem Gardiner.Ito.itoSum_unif_instance (n : ℕ) (hn : 0 < n) :
            ∫ (ω : StdWienerSpace), (itoSum stdWiener (unifPart 0 1 n) n ω - 1 / 2 * (stdWiener 1 ω ^ 2 - stdWiener 0 ω ^ 2 - (1 - 0))) ^ 2 ∂stdWienerMeasure = 1 / (2 * ↑n)

            Instance of (4.2.20): for the standard Wiener process and the uniform partition of [0, 1], ⟨(Σ W(tᵢ₋₁)ΔWᵢ - ½[W(1)² - W(0)² - 1])²⟩ = 1/(2n).

            theorem Gardiner.Ito.qv_msLim_unif {t₀ t : ℝ} (h0 : 0 ≤ t₀) (h : t₀ ≤ t) :
            Filter.Tendsto (fun (k : ℕ) => ∫ (ω : StdWienerSpace), (qvSum stdWiener (unifPart t₀ t (k + 1)) (k + 1) ω - (t - t₀)) ^ 2 ∂stdWienerMeasure) Filter.atTop (nhds 0)

            The uniform partitions realise ms-lim Σ ΔWᵢ² = t - t₀ for the standard Wiener process.

            Erratum (4.2.18) #

            The right-hand side of (4.2.18) as printed: 2 Σᵢ Δtᵢ² + Σ_{i,j} [Δtᵢ - (t - t₀)][Δtⱼ - (t - t₀)].

            Equations
            Instances For

              Erratum (4.2.18). For the uniform partition of [0, 1] into two cells the mean-square error of the quadratic-variation sum is 1, but the printed right-hand side of (4.2.18) gives 2. The correct value is 2 Σ Δtᵢ² (Gardiner.Ito.IsWiener.qv_ms_error).