Omnichannel facility location and fulfillment optimization

E-commerce has been evolving towards omnichannel as we have witnessed many once pure e-retailers opening physical stores or showrooms to enable customers to inspect the product and quickly assess product fit. In omnichannel transportation operations, these stores can strategically fulfill online orders to improve delivery efficiency and reduce transportation costs. Motivated by this trend, this paper studies a pure e-retail company that plans to introduce the offline channel with physical stores and practice omnichannel distribution. For the e-retailer, opening stores can benefit from the reduced transportation cost and the increased revenue, both of which inherently depend on how customers choose channels and stores to make purchases. To this end, we employ a discrete choice model to estimate customers' choices and then propose a modeling framework that jointly optimizes the location of stores and the omnichannel fulfillment strategy to maximize the profit, accounting for the revenue, the transportation cost, and the fixed operating cost of stores. To solve the model of practical scale, we develop a branch-and-cut algorithm, which leverages Benders cuts that are generated on-the-fly using two analytical separation functions. Our computational studies suggest that the proposed algorithm is efficient for solving large-scale problems. Finally, we conduct sensitivity analysis and draw insightful observations.