A survey of dynamic pickup and delivery problems

Abstract

Due to the ubiquitous real-world applications of logistics and supply chain management over the past two decades, dynamic pickup and delivery problems (DPDPs), as a subclass of dynamic vehicle routing problems in which objects or people must be collected and delivered in real time, have become a popular research topic in the field of combinatorial optimization. This article provides a comprehensive review of the DPDP literature from 2004 to 2023, in which their corresponding characteristics, principles, and theoretical analysis are discussed in detail. Furthermore, a taxonomy of the related solution methods for DPDPs is given, which can be segmented into four categories: exact methods, heuristics, metaheuristics, and learning-based methods. Moreover, some experimental comparisons and analysis of up-to-date real-word DPDP benchmarks from Huawei Company are included. Finally, a brief conclusion is given to summarize some potential future directions for DPDPs.