Documentation

SHSLib.Gardiner.Ito.StochDE.Picard

Existence of solutions of Itô SDEs: Picard iteration #

Gardiner §4.3.1, (4.3.10)–(4.3.11). Under the Lipschitz and growth conditions on [t₀, T] (SDECoeffs a b t₀ T K), for every horizon T and every square integrable, 𝓕 t₀-measurable initial value ξ, the Picard iterates x₀ = ξ, x_{n+1}(t) = ξ + ∫_{t₀}^t a(x_n(s), s) ds + ∫_{t₀}^t b(x_n(s), s) dW(s) (taken in continuous adapted versions) converge almost surely uniformly on [t₀, T], and in E sup |·|², to a solution of (4.3.4).

structure Gardiner.Ito.StochDE.IsRegProc {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (t₀ T : ℝ) (x : ℝ → Ω → ℝ) :

A regular process on [t₀, T]: jointly measurable, adapted on [t₀, T], almost surely continuous on [t₀, T], with E sup_{t₀ ≤ t ≤ T} |x(t)|² < ∞.

Instances For
    theorem Gardiner.Ito.StochDE.lintegral_supSq_le {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {P : MeasureTheory.Measure Ω} {Y : ℝ → Ω → ℝ} {t₀ t T : ℝ} (htT : t ≤ T) :
    ∫⁻ (s : ℝ) in Set.Ioc t₀ t, supSq P Y t₀ s ≤ ENNReal.ofReal (t - t₀) * supSq P Y t₀ T
    theorem Gardiner.Ito.StochDE.IsRegProc.sq_int {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {t₀ T : ℝ} {x : ℝ → Ω → ℝ} (hx : IsRegProc 𝓕 P t₀ T x) :
    ∫⁻ (s : ℝ) in Set.Ioc t₀ T, ∫⁻ (ω : Ω), ‖x s ω‖ₑ ^ 2 ∂P < ⊤
    theorem Gardiner.Ito.StochDE.ae_eq_of_supSq_sub_eq_zero {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {P : MeasureTheory.Measure Ω} {x y : ℝ → Ω → ℝ} {t₀ T : ℝ} (htT : t₀ ≤ T) (hxm : Measurable (Function.uncurry x)) (hym : Measurable (Function.uncurry y)) (hxc : ∀ᵐ (ω : Ω) ∂P, ContinuousOn (fun (t : ℝ) => x t ω) (Set.Icc t₀ T)) (hyc : ∀ᵐ (ω : Ω) ∂P, ContinuousOn (fun (t : ℝ) => y t ω) (Set.Icc t₀ T)) (h : supSq P (fun (s : ℝ) (ω : Ω) => x s ω - y s ω) t₀ T = 0) :
    ∀ᵐ (ω : Ω) ∂P, ∀ t ∈ Set.Icc t₀ T, x t ω = y t ω

    Processes with E sup |x - y|² = 0 are indistinguishable.

    theorem Gardiner.Ito.StochDE.supSq_sub_triangle {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {P : MeasureTheory.Measure Ω} {x y z : ℝ → Ω → ℝ} {t₀ T : ℝ} (htT : t₀ ≤ T) (hxm : Measurable (Function.uncurry x)) (hym : Measurable (Function.uncurry y)) (hxc : ∀ᵐ (ω : Ω) ∂P, ContinuousOn (fun (t : ℝ) => x t ω) (Set.Icc t₀ T)) (hyc : ∀ᵐ (ω : Ω) ∂P, ContinuousOn (fun (t : ℝ) => y t ω) (Set.Icc t₀ T)) :
    supSq P (fun (s : ℝ) (ω : Ω) => x s ω - z s ω) t₀ T ≤ 2 * supSq P (fun (s : ℝ) (ω : Ω) => x s ω - y s ω) t₀ T + 2 * supSq P (fun (s : ℝ) (ω : Ω) => z s ω - y s ω) t₀ T

    E sup |x - z|² ≤ 2 E sup |x - y|² + 2 E sup |y - z|².

    Elementary series bounds #

    theorem Gardiner.Ito.StochDE.tsum_sq_le_two_mul (f : ℕ → ENNReal) :
    (∑' (k : ℕ), f k) ^ 2 ≤ 2 * ∑' (k : ℕ), 2 ^ k * f k ^ 2

    A weighted Cauchy–Schwarz inequality: (∑ f k)² ≤ 2 ∑ 2ᵏ (f k)².

    theorem Gardiner.Ito.StochDE.pow_div_factorial_add_le {c : ℝ} (hc : 0 ≤ c) (k n : ℕ) :
    c ^ (k + n) / ↑(k + n).factorial ≤ c ^ k / ↑k.factorial * (c ^ n / ↑n.factorial)
    theorem Gardiner.Ito.StochDE.limUnder_tail_of_tsum_ne_top {f : ℕ → ℝ → ℝ} {s : Set ℝ} (hf : ∀ (j : ℕ), ContinuousOn (f j) s) (hsum : ∑' (j : ℕ), ⨆ u ∈ s, ‖f (j + 1) u - f j u‖ₑ ≠ ⊤) :
    ContinuousOn (fun (x : ℝ) => limUnder Filter.atTop fun (j : ℕ) => f j x) s ∧ ∀ x ∈ s, ∀ (n : ℕ), ‖(limUnder Filter.atTop fun (j : ℕ) => f j x) - f n x‖ₑ ≤ ∑' (k : ℕ), ⨆ u ∈ s, ‖f (k + n + 1) u - f (k + n) u‖ₑ

    Uniform convergence along increments with finite total sup-variation, with a tail bound.

    The Picard step #

    def Gardiner.Ito.StochDE.picardStep {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (W : ℝ → Ω → ℝ) (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (a b : ℝ → ℝ → ℝ) (t₀ T : ℝ) (ξ : Ω → ℝ) (x : ℝ → Ω → ℝ) (t : ℝ) (ω : Ω) :

    The Picard map: x ↦ ξ + ∫_{t₀}^· a(x(s), s) ds + ∫_{t₀}^· b(x(s), s) dW(s), taken in its adapted, continuous version.

    Equations
    Instances For
      def Gardiner.Ito.StochDE.picardSeq {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (W : ℝ → Ω → ℝ) (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (a b : ℝ → ℝ → ℝ) (t₀ T : ℝ) (ξ : Ω → ℝ) :
      ℕ → ℝ → Ω → ℝ

      The Picard iterates x₀ = ξ, x_{n+1} = picardStep x_n.

      Equations
      Instances For
        def Gardiner.Ito.StochDE.sdeSol {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (W : ℝ → Ω → ℝ) (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (a b : ℝ → ℝ → ℝ) (t₀ T : ℝ) (ξ : Ω → ℝ) (t : ℝ) (ω : Ω) :

        The solution of the SDE (4.3.4): the pointwise limit of the Picard iterates.

        Equations
        Instances For
          theorem Gardiner.Ito.StochDE.measurable_xi {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
          theorem Gardiner.Ito.StochDE.IsRegProc.picard {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) {x : ℝ → Ω → ℝ} (hx : IsRegProc 𝓕 P t₀ T x) :
          IsRegProc 𝓕 P t₀ T (picardStep W 𝓕 P a b t₀ T ξ x) ∧ ∀ u ∈ Set.Icc t₀ T, picardStep W 𝓕 P a b t₀ T ξ x u =ᵐ[P] fun (ω : Ω) => (ξ ω + ∫ (s : ℝ) in t₀..u, a (x s ω) s) + itoInt W 𝓕 P u (drive t₀ b x) ω

          The Picard step of a regular process is regular and satisfies the integral equation.

          theorem Gardiner.Ito.StochDE.supSq_picardStep_sub_le {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) {x y : ℝ → Ω → ℝ} (hx : IsRegProc 𝓕 P t₀ T x) (hy : IsRegProc 𝓕 P t₀ T y) (t : ℝ) :
          t ∈ Set.Icc t₀ T → supSq P (fun (s : ℝ) (ω : Ω) => picardStep W 𝓕 P a b t₀ T ξ x s ω - picardStep W 𝓕 P a b t₀ T ξ y s ω) t₀ t ≤ ENNReal.ofReal (sdeRate K (T - t₀)) * ∫⁻ (s : ℝ) in Set.Ioc t₀ t, supSq P (fun (s : ℝ) (ω : Ω) => x s ω - y s ω) t₀ s

          The Picard step is Lipschitz for the E sup |·|² distance (with Gronwall rate).

          theorem Gardiner.Ito.StochDE.isRegProc_const {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
          IsRegProc 𝓕 P t₀ T fun (x : ℝ) (ω : Ω) => ξ ω

          The initial iterate x₀ = ξ is regular.

          theorem Gardiner.Ito.StochDE.isRegProc_picardSeq {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) (n : ℕ) :
          IsRegProc 𝓕 P t₀ T (picardSeq W 𝓕 P a b t₀ T ξ n)

          All Picard iterates are regular.

          Convergence of the Picard iterates #

          def Gardiner.Ito.StochDE.picardGap {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (W : ℝ → Ω → ℝ) (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (a b : ℝ → ℝ → ℝ) (t₀ T : ℝ) (ξ : Ω → ℝ) (n : ℕ) (ω : Ω) :

          sup_{t₀ ≤ u ≤ T} |x_{n+1}(u) - x_n(u)|.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For
            def Gardiner.Ito.StochDE.picardM {Ω : Type u_1} [mΩ : MeasurableSpace Ω] (W : ℝ → Ω → ℝ) (𝓕 : MeasureTheory.Filtration ℝ mΩ) (P : MeasureTheory.Measure Ω) (a b : ℝ → ℝ → ℝ) (t₀ T : ℝ) (ξ : Ω → ℝ) :

            E sup |x₁ - x₀|².

            Equations
            • One or more equations did not get rendered due to their size.
            Instances For
              theorem Gardiner.Ito.StochDE.picardM_lt_top {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              picardM W 𝓕 P a b t₀ T ξ < ⊤
              theorem Gardiner.Ito.StochDE.supSq_picardSeq_succ_sub_le {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) (n : ℕ) (t : ℝ) :
              t ∈ Set.Icc t₀ T → supSq P (fun (s : ℝ) (ω : Ω) => picardSeq W 𝓕 P a b t₀ T ξ (n + 1) s ω - picardSeq W 𝓕 P a b t₀ T ξ n s ω) t₀ t ≤ picardM W 𝓕 P a b t₀ T ξ * gronR (sdeRate K (T - t₀)) t₀ n t

              The Picard bound E sup_{[t₀, t]} |x_{n+1} - x_n|² ≤ M (L (t - t₀))ⁿ / n!.

              theorem Gardiner.Ito.StochDE.aemeasurable_picardGap {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) (n : ℕ) :
              AEMeasurable (picardGap W 𝓕 P a b t₀ T ξ n) P
              theorem Gardiner.Ito.StochDE.lintegral_tsum_picardGap_sq_le {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) (n : ℕ) :
              ∫⁻ (ω : Ω), (∑' (k : ℕ), picardGap W 𝓕 P a b t₀ T ξ (k + n) ω) ^ 2 ∂P ≤ 2 * picardM W 𝓕 P a b t₀ T ξ * ENNReal.ofReal (Real.exp (2 * (sdeRate K (T - t₀) * (T - t₀)))) * ENNReal.ofReal ((sdeRate K (T - t₀) * (T - t₀)) ^ n / ↑n.factorial)

              E (∑_{k ≥ n} sup |x_{k+1} - x_k|)² ≤ 2 M e^{2 L (T - t₀)} (L (T - t₀))ⁿ / n!.

              theorem Gardiner.Ito.StochDE.sdeSol_def {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (t : ℝ) (ω : Ω) :
              sdeSol W 𝓕 P a b t₀ T ξ t ω = limUnder Filter.atTop fun (n : ℕ) => picardSeq W 𝓕 P a b t₀ T ξ n t ω
              theorem Gardiner.Ito.StochDE.ae_sdeSol_spec {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              ∀ᵐ (ω : Ω) ∂P, ContinuousOn (fun (t : ℝ) => sdeSol W 𝓕 P a b t₀ T ξ t ω) (Set.Icc t₀ T) ∧ ∀ t ∈ Set.Icc t₀ T, ∀ (n : ℕ), ‖sdeSol W 𝓕 P a b t₀ T ξ t ω - picardSeq W 𝓕 P a b t₀ T ξ n t ω‖ₑ ≤ ∑' (k : ℕ), picardGap W 𝓕 P a b t₀ T ξ (k + n) ω

              Almost surely, the Picard iterates converge uniformly on [t₀, T] to sdeSol, which is continuous there, with the tail bound |x(t) - x_n(t)| ≤ ∑_{k ≥ n} sup |x_{k+1} - x_k|.

              theorem Gardiner.Ito.StochDE.supSq_sdeSol_sub_le {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) (n : ℕ) :
              supSq P (fun (s : ℝ) (ω : Ω) => sdeSol W 𝓕 P a b t₀ T ξ s ω - picardSeq W 𝓕 P a b t₀ T ξ n s ω) t₀ T ≤ 2 * picardM W 𝓕 P a b t₀ T ξ * ENNReal.ofReal (Real.exp (2 * (sdeRate K (T - t₀) * (T - t₀)))) * ENNReal.ofReal ((sdeRate K (T - t₀) * (T - t₀)) ^ n / ↑n.factorial)

              Mean-square convergence of the Picard iterates: E sup_{[t₀, T]} |x - x_n|² ≤ 2 M e^{2 L (T - t₀)} (L (T - t₀))ⁿ / n!.

              theorem Gardiner.Ito.StochDE.isRegProc_sdeSol {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              IsRegProc 𝓕 P t₀ T (sdeSol W 𝓕 P a b t₀ T ξ)

              The limit of the Picard iterates is a regular process.

              theorem Gardiner.Ito.StochDE.sdeSol_fixed {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              ∀ᵐ (ω : Ω) ∂P, ∀ t ∈ Set.Icc t₀ T, sdeSol W 𝓕 P a b t₀ T ξ t ω = picardStep W 𝓕 P a b t₀ T ξ (sdeSol W 𝓕 P a b t₀ T ξ) t ω

              The limit is a fixed point of the Picard map, up to indistinguishability.

              theorem Gardiner.Ito.StochDE.isSDESol_sdeSol {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              IsSDESol W 𝓕 P a b t₀ T ξ (sdeSol W 𝓕 P a b t₀ T ξ)

              Existence (Gardiner §4.3.1): under the Lipschitz and growth conditions (4.3.10)–(4.3.11) on [t₀, T], for every horizon T ≥ t₀ ≥ 0 and every square integrable, 𝓕 t₀-measurable initial value ξ, the limit sdeSol of the Picard iterates is a continuous adapted solution of (4.3.4) on [t₀, T].

              theorem Gardiner.Ito.StochDE.existsUnique_sdeSol {Ω : Type u_1} [mΩ : MeasurableSpace Ω] {𝓕 : MeasureTheory.Filtration ℝ mΩ} {P : MeasureTheory.Measure Ω} {W : ℝ → Ω → ℝ} {a b : ℝ → ℝ → ℝ} {t₀ T K : ℝ} {ξ : Ω → ℝ} (hW : IsFWiener W 𝓕 P) (hc : SDECoeffs a b t₀ T K) (ht₀ : 0 ≤ t₀) (htT : t₀ ≤ T) (hξm : MeasureTheory.StronglyMeasurable ξ) (hξ : MeasureTheory.MemLp ξ 2 P) :
              ∃ (x : ℝ → Ω → ℝ), IsSDESol W 𝓕 P a b t₀ T ξ x ∧ ∀ (y : ℝ → Ω → ℝ), IsSDESol W 𝓕 P a b t₀ T ξ y → ∀ᵐ (ω : Ω) ∂P, ∀ t ∈ Set.Icc t₀ T, y t ω = x t ω

              Existence and uniqueness (Gardiner §4.3.1): there is a continuous adapted solution of (4.3.4) on [t₀, T], and any two solutions are indistinguishable.