Dynamic lot sizing with stochastic demand timing

In this paper, a novel way of modeling uncertainty on demand in the single-item dynamic lot sizing problem is proposed and studied. The not related to quantity, but rather timing, i.e., fully occurs single period given time interval with probability no partial delivery allowed. first motivated modeled. Our naturally correlates uncertain demands different periods contrary most literature sizing. Dynamic programs are then for general case multiple stochastic timing several special cases. We also show that where backlog cost depends both NP-hard.