Approximation Algorithms for Capacitated Stochastic Lot-sizing Inventory Control Models

Introduction. We study a class of capacitated stochastic lot-sizing inventory control problems with stochastic, non-stationary and correlated demands that evolve over time. Economies of scale and capacity constraints exist in many practical scenarios. However, models with fixed ordering costs and capacity constraints are typically computationally intractable, and even the structure of the optimal policies is not well understood. Thus, computing provably good policies is usually very challenging. In this paper, we develope new provably near-optimal approximation algorithms for a class of core inventory management models with fixed ordering costs and capacity constraints.