Partial automation of the dispatching process to reduce empty kilometers
summary
Project uses real-time ERP and DMS feeds for roughly 2,000 trucks covering ~900,000 km/year and a tree-based recommender that ranks optimal freight-for-truck and truck-for-freight matches, claiming potential savings of about 0.79 kg CO2 per avoided km. Implementation is a Python API (Flask) inside Docker swarm exposed by nginx and uwsgi, integrated into the ERP web UI so dispatchers hit a button, get ranked suggestions, and logging and hardening ensure production stability and traceability.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:38Presentation
- 3:23Background
- 4:32Business perspective
- 7:22Workshop
- 8:27Team
- 9:16Prototyping
- 10:08What have we done
- 11:04Screenshots
- 12:59Integration
- 14:03Work packages
- 14:48Summary
- 16:40Data Science Talk
related talks
33mBuilding Data Science Tools for Sustainable Transformation
20mPartnering with Posit for progress on Environmental Stewardship
32mLean Experimentation and Decision Science with Humans in Warehouses
26mDriving towards a cleaner future through data
34mMove Your Crate - How to optimize your mobility
31mKeeping up with the renewable energy transition
20mOptimising Low-Emissions Zone Geofencing Strategy using SUMO with Vehicle Co-Simulation
50mAnalytics – Balancing ‘Advanced’ and ‘Actionable’