Warehouse Picking: Order batching and routing with time windows

Given a set of customer orders and storage locations, the order batching problems seeks a grouping of the orders into picking routes, such that the overall length of all picking routes is minimized, while respecting the capacity of the picking vehicles. Order picking is estimated to take up more than half of the operational costs of a warehouse. We study a variation of the order batching problem where a time window (release time and due date) is assigned to every order. This is a very common extra constraint in most reallife order batching problems. The proposed solution is based on a GRASP heuristic, where we try to balance proximity of orders against urgency of orders. Computational experiments are performed on simulated data sets from the literature as well as on the data received from a central supermarket chain warehouse in Slovenia. The results of the random data sets allow us to analyze the impact of the time windows on the number of vehicles and the number of batches, while the results on the real data give a significant reduction (up to 75%) on both the number of vehicles, duration of picking (makespan), and the number of batches.