Research question
Which courier should receive each delivery order, and what service fee should the platform offer as demand and courier availability change?
Approach and decision mechanism
The paper formulates joint order assignment and courier service-fee decisions as a Markov decision process. A utility-based choice model captures couriers’ willingness to accept assignments, while a neural value-function approximation accounts for the consequences of current decisions for future operations.
- Evaluate acceptance and future valueUse outstanding orders, courier tasks, and a utility-based choice model to evaluate immediate returns and continuation value.
- Determine pair-specific courier feesUnder a pairwise-separable approximation, select the service fee for each admissible order–courier pair.
- Assign orders and updateSolve the resulting linear assignment problem, observe courier responses, and update decisions as the delivery system evolves.
The service fee is a payment from the platform to the courier, rather than a price charged to the customer. Continuation value helps balance immediate profit against the future use of courier capacity. The two-stage pricing–assignment decomposition applies to the within-epoch problem under the pairwise-separable approximation.
Evidence and scope
Computational experiments use networks constructed from Shenzhen, comparing the proposed method with random, distance-based, and myopic policies. In the expanded 6 km × 6 km scenario with 60 couriers, average platform profit increased by 4.5% relative to the myopic policy; on-time delivery rose from 80.7% to 86.2%, and cancellations fell from 9.3% to 7.7%. Each policy was evaluated over five independent simulation runs.
The results come from simulation on Shenzhen-derived networks. Courier-choice parameters and demand assumptions require calibration for another operation. The pairwise decomposition does not directly accommodate cross-pair pricing constraints such as a global payment budget, and the study considers one platform.
Read the publication recordCite this paper
Yang, C., Chow, A. H. F., Deng, Y., & Zhang, Z. (2027). Dynamic optimization of order assignment and service pricing for urban goods delivery via crowdsourcing with uncertainties. Transportation Research Part E: Logistics and Transportation Review, 218, 105362. https://doi.org/10.1016/j.tre.2026.105362
View BibTeX
@article{yang2027crowdsourced,
title = {Dynamic optimization of order assignment and service pricing for urban goods delivery via crowdsourcing with uncertainties},
author = {Chunwei Yang and Andy H. F. Chow and Yang Deng and Zeyu Zhang},
journal = {Transportation Research Part E: Logistics and Transportation Review},
year = {2027},
volume = {218},
pages = {105362},
doi = {10.1016/j.tre.2026.105362},
url = {https://doi.org/10.1016/j.tre.2026.105362}
}