Dynamic order forecast sharing with dual shipping options under quantity flexibility contract

We study how a manufacturer shares demand information with its supplier under an important class of supply contracts known as quantity flexibility (QF) contract. In this article, we assume a manufacturer, facing evolving demand forecasts, needs to share the order forecasts to the supplier in a rolling-horizon manner. In each period, previous order forecasts can be adjusted but must remain within the flexibility outlined by the QF contract. Additionally, the manufacturer has two shipping options for transportation between the supplier and its location. We formulate this problem using stochastic dynamic programming and solve it with the stochastic dual dynamic programming (SDDP) algorithm. We carry out extensive numerical experiments and demonstrate that this method outperforms the other two benchmark methods. We also show that adopting dual shipping options can substantially reduce the overall costs of this inventory system and improve the effectiveness of information sharing. Most importantly, we find that the symmetric QF contract always leads to the exaggeration of quantities in order forecasts, while the asymmetric QF contract, with increasing flexibility slightly higher than decreasing, can facilitate unbiased order forecast sharing. Therefore, from the supplier's perspective, such asymmetric contracts are preferable.