475 / 2024-04-25 16:18:17

Data-driven Robust Assortment Planning

Assortment planning,Robust optimization

We propose a data-driven approach to robust assortment planning. Based on historical sales data, we build an integrated optimization model that is aimed to estimate customer preferences and yield the optimal assortment simultaneously. Compared to two-stage estimate-then-optimize approaches, our integrated model results in high-quality assortments that are robust to sampling error (i.e., sales data uncertainty). An efficient algorithm is developed to solve the optimization model. We illustrate both theoretically and numerically the robustness of our approach in terms of finding optimal or nearly optimal assortments.