Integral sum weight increments mem Icc
ProbabilityTheory.integral_sum_weight_increments_mem_Icc
Plain-language statement
Two-sided expectation bound for adapted {0,1}-weighted increment sums of X, from the boundedness of elementary stochastic integrals at time t. The lower bound uses the complementary weights 1 - W.
Exact Lean statement
lemma integral_sum_weight_increments_mem_Icc [OrderBot ι]
(hXint : ∀ s, Integrable (X s) μ) {C : ℝ}
(hC : ∀ S : ElementaryPredictableSet 𝓕, μ[(S.indicator (1 : ℝ) ● X) t] ≤ C)
{n : ℕ} {idx : ℕ → ι} (hidx : Monotone idx) (hidxt : ∀ k, idx k ≤ t)
{W : ℕ → Ω → ℝ} (hW01 : ∀ k, k < n → ∀ ω, W k ω = 0 ∨ W k ω = 1)
(hWmeas : ∀ k, k < n → Measurable[𝓕 (idx k)] (W k)) :
∫ ω, (∑ k ∈ range n, W k ω * (X (idx (k + 1)) ω - X (idx k) ω)) ∂μ
∈ Set.Icc (-(C + ∫ ω, |X (idx n) ω - X (idx 0) ω| ∂μ)) CFormal artifact
Lean source
lemma integral_sum_weight_increments_mem_Icc [OrderBot ι] (hXint : ∀ s, Integrable (X s) μ) {C : ℝ} (hC : ∀ S : ElementaryPredictableSet 𝓕, μ[(S.indicator (1 : ℝ) ● X) t] ≤ C) {n : ℕ} {idx : ℕ → ι} (hidx : Monotone idx) (hidxt : ∀ k, idx k ≤ t) {W : ℕ → Ω → ℝ} (hW01 : ∀ k, k < n → ∀ ω, W k ω = 0 ∨ W k ω = 1) (hWmeas : ∀ k, k < n → Measurable[𝓕 (idx k)] (W k)) : ∫ ω, (∑ k ∈ range n, W k ω * (X (idx (k + 1)) ω - X (idx k) ω)) ∂μ ∈ Set.Icc (-(C + ∫ ω, |X (idx n) ω - X (idx 0) ω| ∂μ)) C := by have hb01 : ∀ (V : ℕ → Ω → ℝ), (∀ k, k < n → ∀ ω, V k ω = 0 ∨ V k ω = 1) → ∀ k, k < n → ∀ ω, ‖V k ω‖ ≤ 1 := by intro V hV k hk ω rcases hV k hk ω with h | h <;> simp [h] have hupper : ∀ (V : ℕ → Ω → ℝ), (∀ k, k < n → ∀ ω, V k ω = 0 ∨ V k ω = 1) → (∀ k, k < n → Measurable[𝓕 (idx k)] (V k)) → ∫ ω, (∑ k ∈ range n, V k ω * (X (idx (k + 1)) ω - X (idx k) ω)) ∂μ ≤ C := by intro V hV01 hVmeas let S := elemPredSetOfSeq hidx n hVmeas have hS := integral_elemPredSetOfSeq (𝓕 := 𝓕) (X := X) n hidx hidxt (W := V) hV01 hVmeas rw [← funext_iff] at hS rw [← hS] exact hC S constructor · -- lower bound via the complementary weights have hW1 : ∀ k, k < n → ∀ ω, (1 - W k ω) = 0 ∨ (1 - W k ω) = 1 := by grind have hW1meas : ∀ k, k < n → Measurable[𝓕 (idx k)] (fun ω ↦ 1 - W k ω) := fun k hk ↦ (measurable_const.sub (hWmeas k hk)) have hco := hupper (fun k ω ↦ 1 - W k ω) hW1 hW1meas have hsplit : ∀ ω, (∑ k ∈ range n, W k ω * (X (idx (k + 1)) ω - X (idx k) ω)) = (X (idx n) ω - X (idx 0) ω) - ∑ k ∈ range n, (1 - W k ω) * (X (idx (k + 1)) ω - X (idx k) ω) := by intro ω have htel : ∑ k ∈ range n, (X (idx (k + 1)) ω - X (idx k) ω) = X (idx n) ω - X (idx 0) ω := sum_range_sub (fun k ↦ X (idx k) ω) n rw [← htel, ← sum_sub_distrib] grind have hint1 : Integrable (fun ω ↦ ∑ k ∈ range n, (1 - W k ω) * (X (idx (k + 1)) ω - X (idx k) ω)) μ := integrable_sum_weight_increments hXint (fun k hk ↦ ((hW1meas k hk).mono (𝓕.le _) le_rfl).aestronglyMeasurable) (hb01 _ hW1) have hintsub : Integrable (fun ω ↦ X (idx n) ω - X (idx 0) ω) μ := (hXint _).sub (hXint _) have hI : ∫ ω, (∑ k ∈ range n, W k ω * (X (idx (k + 1)) ω - X (idx k) ω)) ∂μ = (∫ ω, (X (idx n) ω - X (idx 0) ω) ∂μ) - ∫ ω, (∑ k ∈ range n, (1 - W k ω) * (X (idx (k + 1)) ω - X (idx k) ω)) ∂μ := by rw [integral_congr_ae (μ := μ) (ae_of_all _ hsplit), integral_sub hintsub hint1] have hT : -(∫ ω, |X (idx n) ω - X (idx 0) ω| ∂μ) ≤ ∫ ω, (X (idx n) ω - X (idx 0) ω) ∂μ := by rw [neg_le] calc -∫ ω, (X (idx n) ω - X (idx 0) ω) ∂μ ≤ |∫ ω, (X (idx n) ω - X (idx 0) ω) ∂μ| := neg_le_abs _ _ ≤ ∫ ω, |X (idx n) ω - X (idx 0) ω| ∂μ := abs_integral_le_integral_abs rw [hI] linarith [hco] · exact hupper W hW01 hWmeas- Project
- Brownian motion
- License
- Apache-2.0
- Commit
- 5077304f5e73
- Source
- BrownianMotion/StochasticIntegral/Quasimartingale/MaximalInequality.lean:489-543
Reuse this declaration
Bring the exact result into your workflow
The import identifies the source module. Your project still needs the pinned package dependency shown on this page.
What this badge means
This completion status comes from the project or community source. It has not yet been represented here as an independent rebuild and axiom audit.
Continue in this project
Related declarations
Continuous Within At Iio indicator Ioc
continuousWithinAt_Iio_indicator_Ioc
Plain-language statement
The indicator of a half-open interval Ioc a b with constant value c is left-continuous: when approached from the left it is eventually constant, so it is continuous within Iio t at t for every t.
Source project: Brownian motion
Person-level attribution pending.
Dense comap val nhds Within Iio ne Bot
Dense.comap_val_nhdsWithin_Iio_neBot
Plain-language statement
This is the dual of Dense.comap_val_nhdsWithin_Ioi_neBot.
Source project: Brownian motion
Person-level attribution pending.
Dense comap val nhds Within Ioi ne Bot
Dense.comap_val_nhdsWithin_Ioi_neBot
Project documentation
This is an auxillary lemma used to prove Dense.monotone_of_isRightContinuous. It is saying that if D is a dense set and a, b are two points such that a < b, then the comap of 𝓝[Set.Ioi a] a under the inclusion D → α is nontrivial. Note that a < b is necessary as this is clearly not true if a is a top element.
Source project: Brownian motion
Person-level attribution pending.