Proof of Proposition 3, p. 25 — edge-by-edge decomposition φ(T, ℓ) = Σ_{jk∈T} ℓ(jk) φ(T, ℓ_k)
ProvedSeqLoss.TreeRules.edge_decompositioncost-sharingliabilityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1trees
Let be a rule on loss trees that satisfies tree additivity. For an edge write for the loss that is on and everywhere else. Then for every tree and loss ,
The page writes for the unit loss on the edge . The decomposition reduces every problem to problems with a single unit loss.
Formalization Note The loss is defined on all of and need not vanish off ; the identity holds as stated, because a rule assigns zero to every agent for losses that vanish on (balance and non-negativity).
Preamble
import Mathlib import Definitions.Def_SeqLoss_TreeRules_Setting import Definitions.Def_SeqLoss_TreeRules_Axioms
Formal statement
namespace SeqLoss.TreeRules
theorem edge_decomposition (φ : Finset (ℕ × ℕ) → (ℕ × ℕ → ℝ) → ℕ → ℝ)
(hφ : IsTreeRule φ) (hadd : TreeAdditive φ) (T : Finset (ℕ × ℕ)) (hT : IsTree T)
(ℓ : ℕ × ℕ → ℝ) (hℓ : IsTreeLoss ℓ) :
∀ i ∈ N T, φ T ℓ i = ∑ e ∈ T, ℓ e * φ T (Pi.single e 1) i := by sorry
end SeqLoss.TreeRules
Source
Gudmundsson, Hougaard & Ko, Sharing Sequentially Triggered Losses, accepted manuscript (2023), p. 25, Appendix A, proof of Proposition 3, first display
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.