An optimal (SSP) solution may serve every customer by her nearest operating open facility
ProvedReliableFacilityLoc.Scenario.ssp_nearest_normalizationfacility-locationp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1reliabilitystochastic-programming
Let be an optimal solution of (SSP). Then there is an assignment such that is also optimal for (SSP) and, for every customer , scenario and facility ,
with the conventions and . In words: each customer is always served by her closest open, operating facility, and ties are broken in favour of the lowest index.
This is the "without loss of generality" step of the second half of the proof of Proposition 1.
Formalization Note The "without loss of generality" is stated as the existence of such a for the same location vector . It relies on ; no sign is needed on or .
Preamble
import Definitions.Def_ReliableFacilityLoc_Scenario_SSP open Finset
Formal statement
namespace ReliableFacilityLoc.Scenario
/-- **Nearest-open-facility normalization of (SSP)** (Cui–Ouyang–Shen, UCTC-FR-2010-02 (Feb. 2010),
A.1, proof of Proposition 1, p. 33 (PDF 35)): "Without loss of generality, we assume that
`Y_ijω = 1` if and only if `j = min{0 ≤ k ≤ J : δ_kω X_k = 1, d_ik ≤ d_ik' ∀ k' ≠ k s.t.
δ_k'ω X_k' = 1}` (by convention, we assume `X_J = 1`)". For every optimal `(X, Y)` of (SSP) there is
an optimal `(X, Y')` of (SSP), with the same `X`, in which every customer in every scenario is
served by her closest operating open facility, ties broken by the lowest index.
Formalization Note: the "without loss of generality" is stated as the existence of such a `Y'`
for the same `X`. It uses `λ_i ≥ 0` (part of `Instance`); no sign is needed on `d` or `φ`. -/
theorem ssp_nearest_normalization {I J R : ℕ} (D : Instance I J R) (X : Fin J → ℝ)
(Y : Fin I → Fin (J + 1) → FailureScenario J → ℝ) (h : D.IsSSPOptimal X Y) :
∃ Y' : Fin I → Fin (J + 1) → FailureScenario J → ℝ,
D.IsSSPOptimal X Y' ∧ ∀ i ω k, Y' i k ω = 1 ↔ D.IsNearestServer X i ω k := by sorry
end ReliableFacilityLoc.Scenario
Source
Cui, Ouyang, Shen, Reliable Facility Location Design under the Risk of Disruptions, UCTC-FR-2010-02 (Feb. 2010), Appendix A.1 (proof of Proposition 1), p. 33 (PDF 35), "Without loss of generality, we assume that Y_ijω = 1 if and only if …"
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.