Documentation

SHSLib.Core.DavisSegmentConditionalRandomWindow

The two segment laws at a history-measurable start age #

The conditional laws of the active and of the completed part of a segment contribution were proved for the start age u = s − T_N produced by a deterministic observation time s. Two features of that situation were used: the segment has already started at s, and the age window (u, u+T] the segment must complete in has the deterministic length T = t − s.

Neither is available when the segment is observed at the random time at which it starts — the situation the remaining field of the conditional step law, the centering of a segment not yet started, leads to. This module therefore proves both laws for an arbitrary history-measurable start age u and, for the completed part, an arbitrary bounded history-measurable window length V, with the two features replaced by the minimal hypotheses that make the statements meaningful:

For the active part nothing but the first is needed, since the weighted survivor identity already accepts random thresholds. For the completed part the density step is the one of the previous module, at a random window length.

The deterministic-observation-time laws are the special case u = s − T_N, V = t − s.

The active part #

theorem SHS.DavisJumpNoise.ConditionalHoldingTimeLaw.setIntegral_segmentActivePart_startAge {Q : Type u_1} [MeasurableSpace Q] [MeasurableSingletonClass Q] {n m : ℕ} {S : SHS Q n m} {hflow : S.HasGlobalFlow} {hf : S.DavisFlowRegular hflow} {hr : S.DavisResetRegular} {Ω : Type u_2} [mΩ : MeasurableSpace Ω] {P : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure P] {ξ : DavisJumpNoise m Ω} {hξ : ξ.MeasurableNoise} {initial : DavisState Q n} {g : DavisState Q n → ℝ} (hlaw : ConditionalHoldingTimeLaw S hflow hf hr P ξ hξ initial) (hcont : S.DavisHazardContinuousAlongFlow hflow) (hreg : S.DavisResetOccupationRegular hflow g) (N : ℕ) (t : NNReal) (uu : Ω → NNReal) (huumeas : Measurable uu) (huv : ∀ (ω : Ω), uu ω ≤ t - ξ.jumpTime N ω) (B : Set Ω) (hB : MeasurableSet B) (hBt : ∀ ω ∈ B, ξ.jumpTime N ω ≤ t) (K : ℝ) (hK : ∀ (ω : Ω), |S.davisSegmentResetOccupationCompensator hflow g (S.davisStateAfterJump hflow ξ initial N ω) (t - ξ.jumpTime N ω)| ≤ K) :
∫ (ω : Ω) in B ∩ {ω : Ω | uu ω < ξ.holdingTime N ω}, S.davisSegmentActivePart hflow ξ initial g N t ω ∂P = ∫ (ω : Ω) in B ∩ {ω : Ω | uu ω < ξ.holdingTime N ω}, S.davisSegmentConditionalActiveMean hflow g (S.davisStateAfterJump hflow ξ initial N ω) (uu ω) (t - ξ.jumpTime N ω) ∂P

The conditional law of the active part at a history-measurable start age. On a pre-jump history event at index N on which segment N has started by t, the mean of the active part at t, restricted to the event that the segment survives the start age u, is the mean of the conditional active mean from u to the age reached at t.

The completed part #

theorem SHS.DavisJumpNoise.setIntegral_segmentCompletedPart_startAge {Q : Type u_1} [MeasurableSpace Q] [MeasurableSingletonClass Q] {n m : ℕ} {S : SHS Q n m} {hflow : S.HasGlobalFlow} {hf : S.DavisFlowRegular hflow} {hr : S.DavisResetRegular} [S.IntensitiesMeasurable] {Ω : Type u_2} [mΩ : MeasurableSpace Ω] {P : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure P] {ξ : DavisJumpNoise m Ω} {hξ : ξ.MeasurableNoise} {initial : DavisState Q n} {g : DavisState Q n → ℝ} (hlaw : ConditionalHoldingTimeLaw S hflow hf hr P ξ hξ initial) (hchan : ConditionalChannelLaw S hflow P ξ initial) (hcont : S.DavisHazardContinuousAlongFlow hflow) (hsurv : Measurable fun (p : DavisState Q n × NNReal) => S.davisSegmentSurvivor hflow p.1 p.2) (hg : Measurable g) (Cg : ℝ) (hgC : ∀ (z : DavisState Q n), |g z| ≤ Cg) (hPhimeas : Measurable fun (p : DavisState Q n × ℝ) => S.davisSegmentCenteredResetProfile hflow g p.1 p.2) (hPhicont : ∀ (z : DavisState Q n), Continuous (S.davisSegmentCenteredResetProfile hflow g z)) (L : ℝ) (hL : 0 ≤ L) (hlip : ∀ (z : DavisState Q n) (r r' : ℝ), |S.davisSegmentCenteredResetProfile hflow g z r - S.davisSegmentCenteredResetProfile hflow g z r'| ≤ L * |r - r'|) (Rmax K : ℝ) (hK : ∀ (z : DavisState Q n), ∀ r ≤ Rmax, |S.davisSegmentCenteredResetProfile hflow g z r| ≤ K) (N : ℕ) (t : NNReal) (hRmax : ↑t ≤ Rmax) (uu V : Ω → NNReal) (huumeas : Measurable uu) (hVmeasF : Measurable V) (Tmax : NNReal) (hVle : ∀ (ω : Ω), V ω ≤ Tmax) (B : Set Ω) (hB : MeasurableSet B) (hBt : ∀ ω ∈ B, ξ.jumpTime N ω ≤ t) (hwin : ∀ ω ∈ B, t - ξ.jumpTime N ω = uu ω + V ω) :
∫ (ω : Ω) in B ∩ {ω : Ω | uu ω < ξ.holdingTime N ω}, S.davisSegmentCompletedPart hflow ξ initial g N t ω ∂P = ∫ (ω : Ω) in B ∩ {ω : Ω | uu ω < ξ.holdingTime N ω}, S.davisSegmentConditionalCompletedMean hflow g (S.davisStateAfterJump hflow ξ initial N ω) (uu ω) (t - ξ.jumpTime N ω) ∂P

The conditional law of the completed part at a history-measurable start age and window length. On a pre-jump history event at index N on which segment N has started by t and on which the age reached at t is u + V, the mean of the completed part at t, restricted to the event that the segment survives the start age u, is the mean of the conditional completed mean over the window.