Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Euler–Maruyama one-step diffusion error: E[Mk2]≤c h(ek+h)E[M_k^2]\le c\,h(e_k+h)E[Mk2​]≤ch(ek​+h) and E[DkMk]=0E[D_kM_k]=0E[Dk​Mk​]=0

Open
AppliedSDE.SarkkaSolin.euler_maruyama_diffusion_local_error

by Himan · Nov 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

probabilitystochastic-analysis

Setting. Let WWW be a Brownian motion on a probability space (Ω,P)(\Omega,P)(Ω,P) with a filtration F\mathcal FF to which WWW is adapted and whose future increments are independent of the past. Let b,σ:R→Rb,\sigma:\mathbb R\to\mathbb Rb,σ:R→R be globally Lipschitz with a common constant LLL, and let T>0T>0T>0. Let XXX be a process with almost surely continuous paths, with E[Xt2]<∞E[X_t^2]<\inftyE[Xt2​]<∞ for t∈[0,T]t\in[0,T]t∈[0,T], such that H(ω,s)=σ(Xs(ω))H(\omega,s)=\sigma(X_s(\omega))H(ω,s)=σ(Xs​(ω)) is jointly measurable, progressively measurable and square integrable on Ω×[0,t]\Omega\times[0,t]Ω×[0,t] for every t>0t>0t>0, and such that for every t∈[0,T]t\in[0,T]t∈[0,T], almost surely,

Xt=x0+∫0tb(Xs) ds+∫0tσ(Xs) dWs,X_t = x_0 + \int_0^t b(X_s)\,ds + \int_0^t \sigma(X_s)\,dW_s,Xt​=x0​+∫0t​b(Xs​)ds+∫0t​σ(Xs​)dWs​,

where the last term is the platform's L2L^2L2 Itô integral stochasticIntegralBrownian. For each n≥1n\ge 1n≥1 let E0n,…,EnnE^n_0,\dots,E^n_nE0n​,…,Enn​ be the Euler–Maruyama scheme on the grid tk=kT/nt_k=kT/ntk​=kT/n with step h=T/nh=T/nh=T/n: E0n=x0E^n_0=x_0E0n​=x0​ and

Ek+1n=Ekn+b(Ekn) h+σ(Ekn) (Wtk+1−Wtk).E^n_{k+1}=E^n_k+b(E^n_k)\,h+\sigma(E^n_k)\,(W_{t_{k+1}}-W_{t_k}).Ek+1n​=Ekn​+b(Ekn​)h+σ(Ekn​)(Wtk+1​​−Wtk​​).

Write It=∫0tσ(Xs) dWsI_t=\int_0^t\sigma(X_s)\,dW_sIt​=∫0t​σ(Xs​)dWs​ (the binder I with hI names this process) and ek=E[(Xtk−Ekn)2]e_k=E[(X_{t_k}-E^n_k)^2]ek​=E[(Xtk​​−Ekn​)2].

Claim. There is a constant c≥0c\ge 0c≥0, independent of nnn and kkk, such that for all n≥1n\ge 1n≥1 and 0≤k<n0\le k<n0≤k<n the martingale part of the one-step error

Mk=(Itk+1−Itk)−σ(Ekn) (Wtk+1−Wtk)M_k=(I_{t_{k+1}}-I_{t_k})-\sigma(E^n_k)\,(W_{t_{k+1}}-W_{t_k})Mk​=(Itk+1​​−Itk​​)−σ(Ekn​)(Wtk+1​​−Wtk​​)

satisfies

E[Mk2]≤c h (ek+h)andE[(Xtk−Ekn) Mk]=0.E[M_k^2]\le c\,h\,(e_k+h)\qquad\text{and}\qquad E\bigl[(X_{t_k}-E^n_k)\,M_k\bigr]=0.E[Mk2​]≤ch(ek​+h)andE[(Xtk​​−Ekn​)Mk​]=0.

Proof idea. MkM_kMk​ is the Itô integral of (σ(Xs)−σ(Ekn))1(tk,tk+1](s)(\sigma(X_s)-\sigma(E^n_k))\mathbf 1_{(t_k,t_{k+1}]}(s)(σ(Xs​)−σ(Ekn​))1(tk​,tk+1​]​(s), so the Itô isometry gives E[Mk2]=∫tktk+1E[(σ(Xs)−σ(Ekn))2]ds≤2L2(h ek+∫tktk+1E[(Xs−Xtk)2]ds)E[M_k^2]=\int_{t_k}^{t_{k+1}}E[(\sigma(X_s)-\sigma(E^n_k))^2]ds\le 2L^2\bigl(h\,e_k+\int_{t_k}^{t_{k+1}}E[(X_s-X_{t_k})^2]ds\bigr)E[Mk2​]=∫tk​tk+1​​E[(σ(Xs​)−σ(Ekn​))2]ds≤2L2(hek​+∫tk​tk+1​​E[(Xs​−Xtk​​)2]ds), and the increment bound E[(Xs−Xtk)2]≤K(s−tk)E[(X_s-X_{t_k})^2]\le K(s-t_k)E[(Xs​−Xtk​​)2]≤K(s−tk​) finishes the estimate. The orthogonality holds because MkM_kMk​ is a martingale increment over (tk,tk+1](t_k,t_{k+1}](tk​,tk+1​] and Xtk−EknX_{t_k}-E^n_kXtk​​−Ekn​ is square integrable and almost surely equal to an Ftk\mathcal F_{t_k}Ftk​​-measurable variable.

Preamble
import Mathlib
import Definitions.Def_KLSModular_LevyStochCalc_Brownian_Filtered
import Definitions.Def_KLSModular_LevyStochCalc_Brownian_ItoL2CompletionMaster
open MeasureTheory ProbabilityTheory Filter LevyStochCalc.Brownian
open scoped ENNReal NNReal Topology BigOperators
Formal statement
theorem AppliedSDE.SarkkaSolin.euler_maruyama_diffusion_local_error
    {Ω : Type*} [MeasurableSpace Ω]
    (P : Measure Ω) [IsProbabilityMeasure P]
    (W : LevyStochCalc.Brownian.BrownianMotion P)
    (ℱ : Filtration ℝ ‹MeasurableSpace Ω›)
    (hW : LevyStochCalc.Brownian.IsBrownianFiltration W ℱ)
    (T x₀ : ℝ) (hT : 0 < T)
    (b σ : ℝ → ℝ)
    (hLip : ∃ L : ℝ, 0 ≤ L ∧
      (∀ x y : ℝ, |b x - b y| ≤ L * |x - y|) ∧
      (∀ x y : ℝ, |σ x - σ y| ≤ L * |x - y|))
    (X : ℝ → Ω → ℝ)
    (hXcont : ∀ᵐ ω ∂P, Continuous (fun t : ℝ => X t ω))
    (hXmom : ∀ t : ℝ, t ∈ Set.Icc (0 : ℝ) T →
      Integrable (fun ω : Ω => (X t ω) ^ 2) P)
    (H : Ω → ℝ → ℝ)
    (hH : ∀ (ω : Ω) (s : ℝ), H ω s = σ (X s ω))
    (hHmeas : Measurable (Function.uncurry H))
    (hHprog : LevyStochCalc.Probability.ProgressivelyMeasurable ℱ H)
    (hHsq : ∀ (t : ℝ), 0 < t →
      ∫⁻ ω, ∫⁻ s in Set.Icc (0 : ℝ) t,
        (‖H ω s‖₊ : ℝ≥0∞) ^ 2 ∂volume ∂P < ⊤)
    (hXeq : ∀ t : ℝ, 0 ≤ t → t ≤ T →
      (fun ω : Ω => X t ω) =ᵐ[P] fun ω : Ω =>
        x₀ + (∫ s in (0 : ℝ)..t, b (X s ω)) +
          LevyStochCalc.Brownian.Ito.stochasticIntegralBrownian
            W ℱ hW H hHmeas hHprog hHsq t ω)
    (I : ℝ → Ω → ℝ)
    (hI : ∀ (t : ℝ) (ω : Ω), I t ω =
      LevyStochCalc.Brownian.Ito.stochasticIntegralBrownian
        W ℱ hW H hHmeas hHprog hHsq t ω)
    (E : (n : ℕ) → Fin (n + 1) → Ω → ℝ)
    (hE0 : ∀ (n : ℕ) (ω : Ω), E n 0 ω = x₀)
    (hErec : ∀ (n : ℕ), 0 < n →
      ∀ (k : Fin n) (ω : Ω),
        E n k.succ ω = E n k.castSucc ω +
          b (E n k.castSucc ω) * (T / n) +
          σ (E n k.castSucc ω) *
          (W.W ((((k.val + 1 : ℕ) : ℝ) * T / n)) ω -
            W.W (((k.val : ℝ) * T / n)) ω)) :
    ∃ c : ℝ, 0 ≤ c ∧ ∀ (n : ℕ), 0 < n → ∀ k : Fin n,
      (∫ ω, ((I (((k.val + 1 : ℕ) : ℝ) * T / n) ω - I ((k.val : ℝ) * T / n) ω) -
          σ (E n k.castSucc ω) *
            (W.W (((k.val + 1 : ℕ) : ℝ) * T / n) ω - W.W ((k.val : ℝ) * T / n) ω)) ^ 2 ∂P
        ≤ c * (T / n) * ((∫ ω, (X ((k.val : ℝ) * T / n) ω - E n k.castSucc ω) ^ 2 ∂P) + T / n)) ∧
      (∫ ω, (X ((k.val : ℝ) * T / n) ω - E n k.castSucc ω) *
          ((I (((k.val + 1 : ℕ) : ℝ) * T / n) ω - I ((k.val : ℝ) * T / n) ω) -
            σ (E n k.castSucc ω) *
              (W.W (((k.val + 1 : ℕ) : ℝ) * T / n) ω - W.W ((k.val : ℝ) * T / n) ω)) ∂P = 0) := by
  sorry
Source
Särkkä and Solin, Applied Stochastic Differential Equations (Cambridge University Press, 2019), https://doi.org/10.1017/9781108186735, Chapter 8, Algorithm 8.1 and Section 8.2 (Euler–Maruyama method and its strong convergence). Child lemma of the Prove2Me theorem AppliedSDE.SarkkaSolin.euler_maruyama_grid_l2_local_error.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me