Optimization Framework for Crowd-Sourced Delivery Services With the Consideration of Shippers’ Acceptance Uncertainties

Shixuan Hou, Jie Gao, Chun Wang

Research output: Contribution to journalArticleScientificpeer-review


Crowd-Sourced Delivery Services (CDS) use in-store customers, as crowd-shippers, to deliver online orders directly to other customers. As independent contractors, the crowd-shippers are free to decide whether to accept or reject the online orders assigned by the retailer. High order rejection rates can significantly influence the retailer’s delivery costs due to frequent reassignments and shifting the orders to more expensive professional fleet. To incentivize crowd-shippers to accept the matched orders, in this work, we propose a two-stage optimization framework that integrates bipartite matching with an individual compensation scheme. The first stage of the optimization framework computes the optimal matching between crowd-shippers and online orders to minimize the delivery detours and unassigned orders. Given the matching solutions as inputs, the second stage computes personal compensation for each crowd-shipper based on the characteristic of the matched order and his or her acceptance behavior uncertainty, with the goal of minimizing the expected total delivery cost of the retailer. Numerical experiments are conducted using the survey data to illustrate the performance of the proposed framework and compare it with existing matching and pricing strategies in the literature. Our results show that the proposed framework reduces the delivery cost by up to more than 15% and reduces the crowd-shippers’ rejection rate by an average of 55%.
Original languageEnglish
Pages (from-to)684-693
Number of pages10
JournalIEEE Transactions on Intelligent Transportation Systems
Issue number1
Publication statusPublished - 2023
Externally publishedYes


Dive into the research topics of 'Optimization Framework for Crowd-Sourced Delivery Services With the Consideration of Shippers’ Acceptance Uncertainties'. Together they form a unique fingerprint.

Cite this