A multi-channel retail store location model considering customer retry purchasing patterns

The rapid growth of e-commerce has driven retailers to establish multiple retail channels to enhance service quality for their customers. Despite retailers’ efforts to serve customers, services may not always be available due to various reasons. In such cases, customers often retry their purchases through their preferred channel. This behavior, along with the complexities of multi-channel retailing, complicates both the structure and costs of last-mile network design. To optimize store locations and costs, this paper proposes a mixed-integer programming (MIP) model for tactical store location planning, considering customer retry purchasing patterns and three channels: ship from store (SFS), buy online and pick up in store (BOPS), and offline shopping (OS). Given that the problem is NP-hard in the strong sense, we develop an iterative two-phase Lagrangian relaxation and granular tabu search heuristic (LR-GTS) to tackle large-scale instances. In each iteration, the LR operator decomposes the model and produces high-quality location schemes, while the GTS operator improves the vehicle routing in the SFS channel. Numerical results demonstrate that our heuristic exhibits strong performance in solving large-scale problems involving 600 customers. Additionally, we apply our model to real-world cases, offering valuable managerial insights derived from the sensitivity analysis results.