Efficient Order-based Scheduling Algorithms for Automated Retrieval System (ARS) in Smart Warehouses

Abstract

A Smart Warehouse (SW) usually employs an automated warehouse management system that integrates artificial intelligence and robotics to automate the warehouse processes. For instance, an Automated Retrieval System (ARS) is used in SW to retrieve products ordered by customers from shelves in order to speed up the product movement. Such a process is important because the product retrieval speed directly affects the subsequent processes, such as packaging and delivery, thus bringing a significant impact on the entire transaction duration between merchants and customers. In general, an order from a customer may contain several items and these items need to be retrieved from the respective shelves of the warehouse before the packaging process can be started. The retrieval delay of an order can then be defined as the duration from the time when the order being input to the system until the time when the last item of the order is retrieved. To reduce the retrieval delay of orders, we combine the integrality of order with the job scheduling of ARS. We introduce the concept of Order Tag and propose two Order-Based scheduling algorithms according to the greedy strategy to reduce the total retrieval delay of orders. Mathematical models are constructed to classify and evaluate the retrieval problems. Both algorithms take polynomial time and the ideal state is that all the retrieval jobs are evenly distributed to stackers. Simulation results demonstrate that these two strategies can reduce the total retrieval delay by approximately 30% comparing to the existing algorithms.