案例库 · 工程与运营 · 技术决策 · 2019-2022
这条还没译成中文，下面是英文原文。
Deutsche Post routes 13,000 parcel districts on couriers' habits
Learned precedence rules from couriers' own history, then optimised routes so ETAs and live tracking became accurate.
Deutsche Post
那一手
Deutsche Post wanted live tracking and precise delivery time windows for German parcels, but those promises are only as good as the predicted order of deliveries in a tour. The planner's real knowledge, which stops must come before others and which streets flow well, lived in the heads of experienced couriers.
The team built a courier-centric routing algorithm that deduces precedence constraints from historical delivery data, combining optimisation with statistics and machine learning. It generates custom routes for individual couriers and a state-of-the-art model that predicts delivery time windows.
The system is used to plan daily routes for roughly 13,000 parcel delivery districts in Germany and powers live tracking and estimated delivery windows for 1.6 million parcels each day. It was a finalist for the 2022 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research.
为什么管用
- It extracted tacit expertise instead of asking couriers to codify it.
- Personalised routes are more likely to be followed, keeping predictions true.
- Accurate ETAs and tracking directly improve customer experience.
- The approach scales to a whole country's delivery network.
可以搬走什么
Tacit expertise can be mined from historical decisions and turned into constraints an optimiser respects, converting how the best couriers work into software every courier gets.
后来呢
The 2022 Wagner Prize special issue of the INFORMS Journal on Applied Analytics (2023) documents the system, and the approach became part of Deutsche Post's national parcel planning stack.
资料来源
- Introduction: 2022 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research
- Human-Centric Parcel Delivery at Deutsche Post with Operations Research and Machine Learning
发现哪里写错了？告诉我们。