§3.2, p. 16 — in any optimal solution of (MIP),
ProvedFastFashion.Approx.mip_optimal_z_eq_gTildeConsider the program (MIP) (13)–(19) with a finite set of stores , unit selling prices , any warehouse value , warehouse inventory , store inventory , rates , period , a nonempty set of major sizes , and nonempty finite tangent index sets . Then in every optimal solution of (MIP), for every store ,
where , , and is the approximation (8) with the tangent sets .
The statement says that the MIP's sales variables represent the approximate expected sales exactly, so that (MIP) maximizes plus the warehouse value.
Formalization Note The hypothesis is added: the paper's is a unit selling price, and for the variable is not pushed to its bound. The statement is conditional on optimality; the existence of an optimal solution is not asserted. The claim is deterministic and uses the same as the goal theorem.
import Mathlib import Definitions.Def_FastFashion_Approx_Tangents import Definitions.Def_FastFashion_Approx_MIP
namespace FastFashion.Approx
/-- §3.2 (Caro–Gallien, p. 16): in any optimal solution of (MIP) (13)–(19), the variable `z_j`
equals the approximate expected sales `g̃_{λ_j}(x_j + I_j)` of (8) at store `j`, for every store. -/
theorem mip_optimal_z_eq_gTilde {S J : Type*} [Fintype S] [DecidableEq S] [Fintype J]
(P : J → ℝ) (hP : ∀ j, 0 < P j) (K : ℝ) (lam : S → J → ℝ) (hlam : ∀ s j, 0 < lam s j)
(T : ℝ) (hT : 0 < T) (Sp : Finset S) (hSp : Sp.Nonempty)
(Nset : S → J → Finset (WithTop ℕ)) (hNset : ∀ s j, (Nset s j).Nonempty)
(W : S → ℕ) (I : S → J → ℕ) (sol : MIPSol S J)
(hopt : MIPOptimal P K lam T Sp Nset W I sol) :
∀ j, sol.z j = gTilde (fun s => lam s j) T Sp hSp (fun s => Nset s j) (fun s => hNset s j)
(fun s => I s j + sol.x s j) := by sorry
end FastFashion.Approx
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.