p.20 — the recursive property
ProvedRiskStopTC.recursive_propertydynamic-programmingoptimal-stoppingp2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1risk-measuresstochastic-programming
Throughout, is a probability space with a filtration , , , and for a fixed . Relations between random variables hold -almost surely.
For any one-step mappings and nested functionals of (5.1) (with ), and for and every random variable ,
With , , and this is the identity , , on p.20.
It reduces the time- objective of the stopping problem to the time- conditional one, which is how clauses (ii) and (iv) of Theorem 5.6 follow from (i).
Differences from the source:
- generalization: ρ_{0,T}(Z_τ) = ρ_{0,m}(ρ_{m,T}(Z_τ)), m = 1, …, T, for the mappings of Theorem 5.6 → ρ_{s,u}(X) = ρ_{s,t}(ρ_{t,u}(X)) for all s ≤ t ≤ u ≤ T, every function X and arbitrary ϱ; reason: the composition identity needs no axiom; s = 0, t = m, u = T, X = Z_τ gives the page's identity
- encoding: ρ_{s,t} for 0 ≤ s < t ≤ T, and ρ_{T,T}(Z_T) ≡ Z_T (footnote 10) → nested ϱ s t, with ρ_{t,t} = id for every t; reason: the paper fixes ρ_{T,T} = id; the identity at other diagonal indices is used only where the statement reads as the paper's at s = t
Preamble
import Mathlib import Definitions.Def_RiskFunctionals_NestedAVaR_Compositions import Definitions.Def_MultiperiodRisk_Bellman_TestSet import Definitions.Def_RiskStopTC_nested open MeasureTheory RiskFunctionals.NestedAVaR
Formal statement
namespace RiskStopTC
/-- **The recursive property of the nested functionals.** Pichler, Liu & Shapiro,
arXiv:1808.10807v3, p.20: "Note that by the recursive property (4.2) we have that
ρ0,T(Zτ) = ρ0,m(ρm,T(Zτ)), m = 1, . . . , T."
For `s ≤ t ≤ u ≤ T` and every `X`: `ρ_{s,u}(X) = ρ_{s,t}(ρ_{t,u}(X))`; with `s = 0`, `t = m`,
`u = T` and `X = Z_τ` this is the displayed identity.
**Formalization Note.** `ρ_{s,t} = nested ϱ s t` with `ρ_{t,t} = id`. The identity holds for
every function `X` and needs no axiom of the mappings. -/
theorem recursive_property {Ω : Type*} {mΩ : MeasurableSpace Ω} (ℱ : Filtration ℕ mΩ) (P : Measure[mΩ] Ω)
[IsProbabilityMeasure P] (p : ENNReal) (hp : 1 ≤ p) (T : ℕ) (hℱ : IsStandardFiltration ℱ T)
(ϱ : ℕ → (Ω → ℝ) → Ω → ℝ) (s t u : ℕ) (hst : s ≤ t) (htu : t ≤ u) (huT : u ≤ T)
(X : Ω → ℝ) :
nested ϱ s u X = nested ϱ s t (nested ϱ t u X) := by sorry
end RiskStopTC
Source
Pichler, Liu & Shapiro, Risk averse stochastic programming: time consistency and optimal stopping, arXiv:1808.10807v3, p.20 ("Note that by the recursive property (4.2) we have that ρ0,T(Zτ) = ρ0,m(ρm,T(Zτ)), m = 1, . . . , T")
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.