Abstract
This study introduces the same-day delivery time-guarantee (SDDTG) problem for supporting online retail. In the SDDTG, orders are placed online and are processed and delivered from a depot to customer locations in the same day. The problem seeks optimal delivery time guarantees to offer customers as they consider making purchases to increase purchase decision likelihood while accounting for delivery-related, supply-side costs that can affect profits. Time guarantees are decision variables rather than parameters (as is typical) and are designed around anticipated customer satisfaction levels and purchase likelihoods. Delivery deadlines are not merely given to customers once they make their purchases, but rather the attractiveness of the offered delivery guarantees affects whether they make their purchases, i.e., whether demand is realized. The problem is conceptualized as a multi-stage, stochastic, mixed-integer program in which uncertainties associated with customer reaction to delivery time guarantee offers and their arrival over time are captured. Given a shrinking horizon over a fixed planning horizon, the multi-stage program is approximated by a series of two-stage programs. A parallelized progressive hedging solution methodology is proposed and insights from its application on a case study. The problem recognizes tradeoffs between meeting promised delivery times and capturing the market.
