Algorithms and Models for Automated Replenishment of Store Shelves – Exploratory Research

Today’s retail store is a hotbed of constant improvement and innovation to provide a seamlessly smooth experience to the customer. With the permeating use of AI/ML and IoT technology, retail chains are moving rapidly to deploy such innovations to their retail stores to improve physical process efficiency and make the customer experience as smooth as possible. Major retailers like Walmart and Ahold Delhaize are experimenting with using robots to automate repetitive tasks like shelf replenishment. There is little research on the issue of automating the replenishment of merchandise inside retail stores during periods of heavy demand for a particular product. In such cases, employees are usually seen scrambling to refill shelves during store operations, thus interfering with customer experience and decreasing physical process efficiency in general. Our goal is to study an approach to automate the process of replenishing products dynamically by leveraging real time data using shelf-stocking robots. Specifically, we study algorithms that dynamically route shelf-stocking robots while minimizing interference with shoppers. The routing algorithms proposed would incorporate congestion awareness to respond to shopper dynamics, thus avoiding interference with shoppers. One key challenge is to model shopper behavior which can be expected to influence the performance of such algorithms. This exploratory research has the potential to influence the way store fronts of the future are designed, enabled by a new class of algorithms for material handling that improves service levels in nextgen storefronts.