Picard iteration (16)–(19): uniformly on compacts, the unique solution of (13)–(14)
ProvedHarrisonReimanRBM.Orthant.picard_convergencecontractionfixed-pointp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1reflected-brownian-motion
Let and let be a nonnegative matrix with zeros on the diagonal and maximal column sum . Fix and let . Define
Then there is a path such that
- (continuous, nondecreasing, ) and for all ;
- every with for all coincides with on ;
- uniformly on compact intervals of (19).
This gives existence and uniqueness for the fixed-point formulation (13)–(14), together with the explicit construction (16)–(18).
Formalization Note The paper works under , printed as the maximal row sum; as in the contraction milestone, the norm that makes a contraction under the row-vector convention is the maximal column sum, which is assumed here. Uniform convergence on compacts is Reiman84.QueueLength.UocTendsto.
Preamble
import Mathlib import Definitions.Def_Reiman84_QueueLength_Paths import Definitions.Def_HarrisonReimanRBM_Orthant_Basic
Formal statement
namespace HarrisonReimanRBM.Orthant
open Reiman84.QueueLength
/-- Proof of Theorem 1, pp. 304–305, (16)–(19): if `Q ≥ 0` has zero diagonal and maximal
column sum `< 1`, then for `x ∈ C_S` the system (13)–(14) (`y ∈ C₀`, `y = π(y)` on
`[0, ∞)`) has a unique solution `y`, and the Picard iterates `y⁰ ≡ 0`, `yⁿ⁺¹ = π(yⁿ)`
converge to `y` uniformly on compact intervals. -/
theorem picard_convergence {K : ℕ} (hK : 0 < K) (Q : Matrix (Fin K) (Fin K) ℝ)
(hQ0 : ∀ i j, 0 ≤ Q i j) (hQd : ∀ j, Q j j = 0) (hcol : maxColSum Q < 1)
(x : ℝ → Fin K → ℝ) (hx : IsCPlus x) :
∃ y : ℝ → Fin K → ℝ, InC0 y ∧ (∀ t, 0 ≤ t → y t = piMap Q x y t) ∧
(∀ y' : ℝ → Fin K → ℝ, InC0 y' → (∀ t, 0 ≤ t → y' t = piMap Q x y' t) →
∀ t, 0 ≤ t → y' t = y t) ∧
UocTendsto (fun n => picardIter Q x n) y := by sorry
end HarrisonReimanRBM.Orthant
Source
Harrison & Reiman, Reflected Brownian Motion on an Orthant, Ann. Probab. 9(2) (1981), pp. 304–305, proof of Theorem 1, (16)–(19)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.