Large deviations-based asymptotics for inventory control in supply chains

摘要: We consider a model of capacitated single-class supply chain consisting tandem production facilities and propose policies in two cases: (a) when each facility has access to its local inventory only, (b) it knowledge the total downstream inventory. In case proposed policy guarantees stockout probabilities at stage stay bounded below given constants (service level constraints). we minimize expected cost subject service constraints. both cases rely upon large deviations asymptotics analytically obtain parameters; this leads huge computational savings compared simulation. Our can accommodate autocorrelated demand processes, critical features modern failure-prone manufacturing systems. demonstrate that detailed distributional information on which is incorporated into asymptotics, control decisions. Some extensions multiclass setting are discussed.