Designing new models and algorithms to improve order picking operations

Order picking has been identified as a crucial factor for the competitiveness of a supply chain because inadequate order picking performance causes customer dissatisfaction and high costs. This dissertation aims at designing new models and algorithms to improve order picking operations and to support managerial decisions on facing current challenges in order picking. First, we study the standard order batching problem (OBP) to optimize the batching of customer orders with the objective of minimizing the total length of order picking tours. We present a mathematical model formulation of the problem and develop a hybrid solution approach of an adaptive large neighborhood search and a tabu search method. In numerical studies, we conduct an extensive comparison of our method to all previously published OBP methods that used standard benchmark sets to investigate their performance. Our hybrid outperforms all comparison methods with respect to average solution quality and runtime. Compared to the state-of-the-art, the hybrid shows the clearest advantages on the larger instances of the existing benchmark sets, which assume a larger number of customer orders and larger capacities of the picking device. Finally, our method is able to solve newly generated large-scale instances with up to 600 customer orders and six items per customer order with reasonable runtimes and convincing scaling behavior and robustness. Next, we address a problem based on a practical case, which is inspired by a warehouse of a German manufacturer of household products. In this warehouse, heavy items are not allowed to be placed on top of light items during picking to prevent damage to the light items. Currently, the case company determines the sequence for retrieving the items from their storage locations by applying a simple S-shape strategy that neglects this precedence constraint. As a result, order pickers place the collected items next to each other in plastic boxes and sort the items respecting the precedence constraint at the end of the order picking process. To avoid this sorting, we propose a picker routing strategy that incorporates the precedence constraint by picking heavy items before light items, and we develop an exact solution method to evaluate the strategy. We assess the performance of our strategy on a dataset provided to us by the manufacturer. We compare our strategy to the strategy used in the warehouse of the case company, and to an exact picker routing approach that does not consider the given precedence constraint. The results clearly demonstrate the convincing performance of our strategy even if we compare our strategy to the exact solution method that neglects the precedence constraint. Last, we investigate a new order picking problem, in which human order pickers of the traditional picker-to-parts setup are supported by automated guided vehicles (AGVs). We introduce two mathematical model formulations of the problem, and we develop a heuristic to solve the NP-hard problem. In numerical studies, we assess the solution quality of the heuristic in comparison to optimal solutions. The results demonstrate the ability of the heuristic in finding high-quality solutions within a negligible computation time. We conduct several computational experiments to investigate the effect of different numbers of AGVs and different traveling and walking speed ratios between AGVs and order pickers on the average total tardiness. The results of our experiments indicate that by adding (or removing) AGVs or by increasing (or decreasing) the AGV speed to adapt to different workloads, a large number of customer orders can be completed until the respective due date.