Since the late 1950s, lot sizing problems have been extensively studied by researchers and practitioners, leading to a significant body of literature. While deterministic lot sizing problems constitute the majority of these studies, dealing with uncertainties has been a challenging task. A significant limitation in stochastic lot sizing is that stochastic demands are usually treated as independent random variables, i.e the correlation between stochastic demands in different periods, which usually exists in real-life settings, is ignored. To build this correlation, Akartunalı and Dauzere-Peres consider that demand quantities are deterministic but their timing is stochastic.

In current work, we extend the work of Akartunalı and Dauzere-Peres to the multi-item lot sizing problem and propose appropriate solution approaches.