An available-to-promise stochastic model for order promising based on dynamic resource reservation policy

Facing uncertain future customer orders, a pull-based available-to-promise (ATP) mechanism will deteriorate the overall profit since it allocates critical resources only to current customer orders. To prevent current less-profitable customer orders from over-consuming critical resources, this study investigates a push-pull based ATP problem with two time stages and three profit margin levels, and develops a dynamic resource reservation policy to maximise the expected total profit. Then, a corresponding push-pull based stochastic ATP model is established with known independent demand distributions, and the optimal reservation level is derived by the genetic algorithm to maximise the expected total profit. Finally, a series of simulation experiments are conducted to reveal the impact of some key factors, and the experiment results provide theoretical guidance and implementation methods for companies to maximise overall profits.