Physical upper bound for (Proposition 2)
ProvedOWMR.fifo_physical_upper_boundThis is the physical upper bound for the deterministic FIFO policy (Proposition 2 of the source).
Consider an OWMR network with retailers that satisfies the standing assumptions, a valid initial state , and admissible choices for the FIFO policy of Definition 1 (canonical optima, best half windows, mirror centers , and the supplier pair ). Let be the pair that retailer operates, and let
be its augmented cost: the long-run average cost of the policy in retailer 's single-location problem with actual setup cost and holding and backlog cost , plus its mirror cost , the minimum over centers of the long-run average of , where and is the retailer's position including outstanding requests. Then
where is the long-run average expected cost from , , and is the optimal average cost of the supplier's single-location problem with aggregate demand, cost , and setup cost .
The supplier is charged one single-location problem at the inflated setup cost , which pays for continuation shipments, and each retailer is charged its own augmented cost. With the retailer comparison and the shared-cost lower bound (Proposition 3), this bound yields Theorem 1.
Formalization Note Costs are extended nonnegative reals. The augmented cost is N.augCost i (c.s i) (c.S i), the sum of ssAvgCost at setup cost and mirrorCost; a retailer with uses base stock ().
import Definitions.Def_owmr_fifo_policy open scoped ENNReal
namespace OWMR
theorem fifo_physical_upper_bound {n : ℕ} (hn : 0 < n) (N : Network n) (hN : N.Standing)
(x : InitState n) (hx : x.Valid) (c : FIFOChoice n) (hc : IsFIFOChoice N c) :
(fifoPolicy N x c).avgCost N x ≤ N.J0 (N.K0 + N.Kmax) + ∑ i, N.augCost i (c.s i) (c.S i) := by sorry
end OWMR