Integrating order picking and vehicle routing decisions in a dynamic e-commerce setting

Abstract

Responding to e-commerce orders as quickly as possible is indispensable for companies competing in the current retailing landscape. After a customer orders online, the requested items should be picked in the warehouse and delivered to the customer's location. Previous research showed the benefits of integrating the order picking and vehicle routing decisions in a static context. To achieve shorter customer response times, we propose multiple new algorithms in this study, able to handle the integrated problem of picking and delivery in a setting with dynamic order arrivals and considering more complex picking policies. We develop and implement four online large neighbourhood search algorithms, differing in their degree of integration between picking and routing decisions. The algorithms are tested on different operational settings. Differences between the algorithms and the impact of the operating context are analysed by use of an ANOVA. The results highlight the importance of integrated decision-making, as well as the large impact of the operating context on the operational efficiency. Furthermore, we demonstrate the environmental impact of stringent customer requests. These insights can be used by companies to optimise their operations.