Omnichannel Inventory Replenishment Optimization

Abstract

We study the multi-period inventory replenishment problem for an omnichannel retailer that aims to optimally position inventory across its warehouses and stores to meet its uncertain omnichannel demand that includes walk-ins at stores as well as online demand that can be fulfilled from any location. We present an end-to-end solution approach that combines demand forecasting and inventory optimization together with a network planner for real-world deployment. We present various inventory management heuristics that aim to maximize system-wide profit margins, while incorporating various operations constraints. Using real-data from a retailer, we present network-based simulations using the different heuristics to understand the impact on various KPIs and compare against the as-is method.