Asymptotic Optimality of Open-Loop Policies in Lost-Sales Inventory Models with Stochastic Lead Times

Lost-sales inventory systems

“…If e > 0 denotes a desired upper bound on the optimality gap, this theorem implies that, as ε is reduced, the minimum lead time required is O log ( (1 / )) e . For integer-valued demand and order quantities (i.e., non-divisible products), Bai et al (2023) show that constant-order policies are typically not asymptotically optimal. Instead, the authors study a bracket policy, which defines a fixed sequence of order quantities ë û r and é ù r such that the average order quantity is exactly r. The authors show that this bracket policy is asymptotically optimal for non-divisible products.…”