Modelling a batch assorting operation for an autonomous cart in a parallel-aisle order assorting system

This study considers an order assorting operation for an autonomous cart in a parallel-aisle order assorting system (OAS). We build a batch assorting model for the autonomous assortment cart and develop two large-scale heuristic algorithms. The autonomous function holds the information of locations and numbers of assortment which allows the batching model to consider the multiple products in a trip (i.e., batching) and to split them into multiple trips (i.e., binning). We formulate an exact batch assorting model using mixed-integer programming (MIP) to minimize the total travel distance of the cart, and develop two greedy heuristic algorithms for large-scale problems based on the First-fit Decreasing binning algorithm and the Seed order batching algorithm. The exact batch assorting model solves the medium-scale problem within a reasonable time. A comparison of the travel distances shows that the exact batch assorting solution dominates heuristic algorithms over the medium-scale operations and that the First-fit Decreasing algorithm outperforms the modified Seed algorithm in reducing the number of trips. We conclude that the autonomous cart can provide additional performance gains by flexible binning and batching, and that the batch assorting algorithms can be improved to support large-scale operations.