Order batch picking optimization under different storage scenarios for e-commerce warehouses

Highlights

• Investigate order batch picking optimization problems from the perspective of storage system.

• Establish the order batch picking optimization model especially considering multi-location storage systems.

• Propose an order batching optimization algorithm package to deal with multi-scenarioproblems.

• Investigate the efficiency of each combination of algorithms and recommend the promising ones.

Abstract

To improve the operational efficiency of e-commerce warehouses, multi-location storage systems which means each stock keeping unit can be stored in multiple locations or a location can contain multiple stock keeping units, have been developed and applied in practice. When orders are picked in a batch, how to select the picking location from storage locations holding the identical stock keeping unit obviously affects how far the pickers must travel to complete the picking tasks. However, few works have systematically studied how to optimize order batch picking from the perspective of different storage systems. This paper formulated the order batch picking optimization problems for three typical storage systems and developed the algorithm package including location interval distance algorithm, location selection algorithm, routing algorithm and order batching algorithm to tackle them. Our work is particularly capable to covering the situation of multi-location storage system. The numerical experiment results show that the performance of the proposed algorithm combinations is satisfactory to solve the problems with different size both in solution quality and computation efficiency. The applicable algorithm combinations used in practice are also recommended by comparative analysis. Our study can provide valuable decision reference to warehouse managers for operating batch picking system especially under multi-location scenarios efficiently.

Introduction

With the growth of e-commerce in recent years, warehouse operations face the challenges of small orders, large assortment of items in a batch, limited space and tight delivery schedules. Storing stock keeping units (SKUs) more reasonable is one effective approach to improve order fulfillment efficiency and cope with this challenge within warehouse. Beside the traditional random storage, class based storage and full turnover based storage strategies, some new storage strategies have been proposed and applied to benefit warehousing operations in practice. To increase the utilization of restricted storage space, different SKUs are allocated to the same location, which is called the mixed-shelves system. On the other hand, to shrink the average distance from anywhere to the threshold, the scattered storage system is proposed, in which an SKU is scattered elaborately all around the warehouse. It intends to shorten order throughput times by reducing the unproductive walking distance.

No matter mix-shelve system or scatter storage, all these strategies refers to the correspondence relationship between location and SKU in essence. From the perspective of this correspondence relationship, we categorize the storage system into three types: (1) 1–1 storage system: Each SKU has a specific location and does not appear in any other locations; (2) 1–n storage system: Each SKU can be stored in more than one location, but only one SKU is stored in each location; (3) n–n storage system: Each SKU can be stored in multiple locations, and each location can hold multiple SKUs. The last two are multi-location storage systems, which have been becoming prevalent in most e-commerce warehouse.

Although enormous existing literatures have focused and studied the batch picking topics as shown in section 2, most of them have not clarified the correspondence relationship between location and SKU or implied 1–1 storage system. Moreover, rare literature has explored the specific optimization issues brought from multi-location storage systems, such as how to select a picking location from several identical alternative locations to reduce the travel distance, etc. Our work aims to close the gap. The prime contribution of this paper is to comprehensively explore the order batch picking optimization problems by model formulation and algorithms developments from the perspective of different storage scenarios. It is able to particularly cover the optimization of batch picking in multi-location storage system beside the traditional 1–1 storage system.

We establish the order batching optimization models for the typical three storage systems mentioned above. Without loss of generality, we propose an order batching optimization algorithm package, which include a location interval distance algorithm, a location selection algorithm, a routing algorithm, and an order batching algorithm. This algorithm package is applicable to solve the aforementioned models. The remainder of this paper is organized as follows. The related literature is summarized in section 2. In section 3, we formulate the order batching problems for the different storage systems. The order batching algorithm package consisting with various nested algorithms is proposed in section 4. The numerical experiments and the results analysis are presented in section 5. Finally, the conclusions are drawn and suggestions for further research are given in section 6.