The split formulation (19a)–(19l) is a relaxation of (RSP)
ProvedReliableFacilityLoc.RRSP.split_relaxes_rspfacility-locationoperations-researchp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1reliability
Let be feasible for the relaxed subproblem (RSP). Then with is feasible for the split formulation (19a)–(19l), and its cost equals the (RSP) objective:
Hence the optimal value of the split formulation is at most the optimal value of (RSP). This is the last step of the proof of Proposition 4.
Formalization Note No assumption on the data is needed. Feasibility and cost are those of the RSP and split-formulation definitions.
Preamble
import Mathlib import Definitions.Def_ReliableFacilityLoc_RRSP_RSP import Definitions.Def_ReliableFacilityLoc_RRSP_Split open Finset
Formal statement
namespace ReliableFacilityLoc.RRSP
/-- **The split formulation (19a)–(19l) is a relaxation of (RSP_i)** (Cui, Ouyang & Shen,
UCTC-FR-2010-02 (Feb. 2010), Appendix A.4, p. 39, PDF 41, last paragraph; unnumbered). Every
feasible solution `(Y, P, W)` of (RSP_i) gives the feasible solution `(Y, Z, P, W)` with `Z = Y` of
the split formulation (this is constraint (19m)), with the same cost: `G(Y, Y, P) = Φ_i`.
Printed: "Since formulation (19a) - (19l) is a relaxation of (RSP), it follows that (RRSP) yields a
lower bound for (RSP)."
Formalization Note: the split formulation is the one of `IsSplitFeasible` / `splitObjective`
(with the objective the proof of Lemma 1 uses, see there). No hypothesis on the data is needed. -/
theorem split_relaxes_rsp {J R : ℕ} (lam : ℝ) (d : Fin J → ℝ) (phi : ℝ) (q : Fin J → ℝ)
(mu : Fin J → ℝ) (Y P W : Fin (J + 1) → Fin (R + 1) → ℝ) (h : IsRSPFeasible q Y P W) :
IsSplitFeasible q Y Y P W ∧
splitObjective lam d phi mu Y W = rspObjective lam d phi mu Y W := by sorry
end ReliableFacilityLoc.RRSP
Source
Cui, Ouyang, Shen, Reliable Facility Location Design under the Risk of Disruptions, UCTC-FR-2010-02 (Feb. 2010), p. 39 (PDF 41), Appendix A.4, last paragraph of the proof of Proposition 4 (unnumbered)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.