Fixed and rolling batching for order picking from multiple carousels
Résumé
The carousel is an important automated storage and retrieval system. Customer orders to be collected from carousels are grouped into batches and the orders of a batch are collected simultaneously. We study two batching methods for picking from multiple carousels: (i) fixed batching, where all orders belonging to a batch are collected before the next batch is started, and (ii) rolling batching, where each time an order belonging to the current batch is completed, it is immediately replaced by a new order. For these methods, the main decision is respectively to allocate orders to different batches and to define a sequence for adding orders into batches. With the objective of minimising the total completion time, we develop models for fixed and rolling batching. Thanks to our collaboration with the company KLS Logistic Systems, we use real data to test and prove the efficiency of our models. With comparison to current batching strategies of real companies and to benchmark heuristics, our models yield a considerable time saving. Our findings also indicate that an increase in order size favours the rolling batching over the fixed batching. The models developed in this paper have been implemented and are currently being used in practice.