integrable integral real
ProvedProbabilityTheory.ElementaryPredictableSet.integrable_integral_realcontext-reviewlevystochcalcsource-import
The source theorem ProbabilityTheory.ElementaryPredictableSet.integrable_integral_real is extracted from BrownianMotion.StochasticIntegral.SimpleProcess. The original Lean proposition is retained verbatim, and its proof is tracked separately for remote certification.
Preamble
import Definitions.Def_BrownianMotion_StochasticIntegral_SimpleProcess_F000
/-
Copyright (c) 2025 Rémy Degenne. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Rémy Degenne, Thomas Zhu
-/
/-! # Simple processes and elementary stochastic integral
## Main definitions
- `ElementaryPredictableSet`: the type of elementary predictable sets
- `SimpleProcess`: the type of simple processes, as a Module over ℝ
- `SimpleProcess.toFun`: simple process interpreted as a stochastic process, with a CoeFun instance
- `SimpleProcess.integral`: elementary stochastic integral
- `SimpleProcess.isPredictable`: simple processes are predictable
## Implementation notes
`SimpleProcess` consists of a value function as a `Finsupp`: `value : ι × ι →₀ Ω → E` and a
value at ⊥: `valueBot : Ω → E`. This allows the definition of operations like addition to be
defined naturally.
However, this makes the function `SimpleProcess.toFun` non-injective, so `SimpleProcess` is not
`FunLike`. In other words, two distinct elements `X Y : SimpleProcess E 𝓕` may produce the same
function `(⇑X : ι → Ω → E) = (⇑Y : ι → Ω → E)`.
There are subtleties that are caused by this: for example, by a nonnegative simple
process, we mean `X : SimpleProcess E 𝓕` with `0 ≤ X.valueBot` and `0 ≤ X.value`, which is not the
same as `0 ≤ (⇑X : ι → Ω → E)`.
Similarly, `ElementaryPredictableSet` is a data type that has a coercion to `Set (ι × Ω)`, but
this coercion is not injective, so it is not `SetLike`. This makes it easy to define the indicator
function of an elementary predictable set as a simple process by mapping respective datas
(which is why it also requires disjoint unions).
## TODO
- Generalize instance variables.
-/
@[expose] public section
open MeasureTheory Finset Filter ContinuousLinearMap
open scoped ENNReal Topology
noncomputable section
namespace ProbabilityTheory
variable {ι Ω E : Type*} [LinearOrder ι] [OrderBot ι] {mΩ : MeasurableSpace Ω}
variable [NormedAddCommGroup E] [mE : MeasurableSpace E]
-- Note: separable space is needed for `ContinuousAdd.measurableMul₂` (which has the wrong name)
variable [NormedSpace ℝ E] [BorelSpace E] [SecondCountableTopology E]
variable {𝓕 : Filtration ι mΩ}
attribute [local measurability]
measurableSet_predictable_singleton_bot_prod
measurableSet_predictable_Ioi_prod
measurableSet_predictable_Ioc_prod
namespace ElementaryPredictableSet
end ElementaryPredictableSet
namespace SimpleProcess
-- todo: replace the condition on W by IsPredictable
section Module
-- TODO: Mathlib's Finset.measurable_prod is wrong because it is currently an exact duplicate of
-- Finset.measurable_fun_sum; we want the following version instead.
end Module
section ToFun
end ToFun
variable {F : Type*} [NormedAddCommGroup F] [NormedSpace ℝ F]
variable {G : Type*} [NormedAddCommGroup G] [NormedSpace ℝ G]
section Integral
variable {B : E →L[ℝ] F →L[ℝ] G}
-- @[simp] lemma integral_top (V : SimpleProcess E 𝓕) (X : ι → Ω → F) (ω : Ω) :
-- integral B V X ⊤ ω = V.value.sum fun p v ↦ B (v ω) (X p.2 ω - X p.1 ω) := by simp [integral]
end Integral
section Assoc
section Bilin
section map₂
variable [MeasurableSpace F] [BorelSpace F] [SecondCountableTopology F]
variable [MeasurableSpace G] [BorelSpace G] [SecondCountableTopology G]
-- Analogous to LinearMap.finsupp_sum_apply and AddMonoidHom.finsuppSum_apply:
end map₂
variable [MeasurableSpace F] [BorelSpace F] [SecondCountableTopology F]
variable {H : Type*} [NormedAddCommGroup H] [NormedSpace ℝ H]
variable {I : Type*} [NormedAddCommGroup I] [NormedSpace ℝ I]
variable {J : Type*} [NormedAddCommGroup J] [NormedSpace ℝ J]
variable [MeasurableSpace J] [BorelSpace J] [SecondCountableTopology J]
end Bilin
section Comp
variable [SecondCountableTopology (F →L[ℝ] G)]
variable [SecondCountableTopology (E →L[ℝ] F)]
variable [SecondCountableTopology (E →L[ℝ] G)]
omit mE [SecondCountableTopology E]
end Comp
section SMul
-- TODO: cor:elemStochIntegral_assoc_real_bilin and lem:elemStochIntegral_assoc
end SMul
end Assoc
end SimpleProcess
section Indicator
namespace ElementaryPredictableSet
variable {F : Type*} [NormedAddCommGroup F] [NormedSpace ℝ F]
variable {G : Type*} [NormedAddCommGroup G] [NormedSpace ℝ G]
open scoped SimpleProcess
Formal statement
theorem _root_.ProbabilityTheory.ElementaryPredictableSet.integrable_integral_real {μ : Measure Ω} {X : ι → Ω → ℝ}
(hXint : ∀ s, Integrable (X s) μ) (S : ElementaryPredictableSet 𝓕) (c : ℝ) (t : ι) :
Integrable (((S.indicator c) ● X) t) μ := by sorry
Source