Projects / AI Innovation Challenge

Optimizing Delivery Routes in LATAM Using AI Planning

Challenge Completed!


Delivery Route Optimization

Background

Urban congestion in cities like Bogotá, Lima, Mexico City, and Rio de Janeiro presents a significant challenge for last-mile logistics. With increasing e-commerce demand and population growth, delivery services face inefficiencies that lead to higher costs, longer delivery times, and a growing carbon footprint.

To address these challenges, Colombian technology company Carryt partnered with Omdena in the Route Optimization Project. This collaboration aimed to leverage advanced tools and data to create scalable solutions for optimizing logistics operations.

Objective

The project aimed to create a powerful route optimization tool using Google’s OR-Tools and OpenStreetMap (OSM) data. The tool was designed to:

  • Reduce delivery times and distances.
  • Minimize fuel consumption and environmental impact.
  • Enhance driver well-being by creating efficient and manageable routes.
  • Provide a scalable, reproducible solution for other urban areas.

Approach

The project unfolded over eight weeks, with 50 AI engineers collaborating globally to tackle this ambitious challenge. Key steps included:

  1. Understanding the Problem:
    • Collaborators analyzed data provided by Carryt, including Shapefiles and OSM files, to identify inefficiencies in delivery routes and understand logistical constraints.
  2. Data Preparation:
    • Teams conducted data wrangling, preprocessing, and exploratory data analysis to identify and resolve quality issues.
    • A special guide was created for handling challenging datasets to ensure robustness in modeling and deployment.
  3. Modeling and Algorithm Development:
    • Various modeling approaches were explored before selecting Google’s OR-Tools, a versatile open-source suite for operations research.
    • The OR-Tool was customized to integrate traffic, vehicle, and regulatory constraints for precise route mapping.
  4. Implementation and Deployment:
    • A high-performance web service was developed using the Flask framework and deployed on AWS to ensure scalability and usability.

Results and Impact

The Route Optimization Project successfully delivered a route optimization tool that:

  • Optimized Routes in Bogotá: Demonstrated reduced travel times and shorter delivery routes, significantly improving efficiency.
  • Scalability: Designed a reproducible framework applicable to other cities, paving the way for broader adoption.
  • Sustainability and Cost Savings: Lowered carbon emissions and reduced fuel costs, contributing to a greener logistics ecosystem.
  • Driver Benefits: Empowered drivers with manageable workloads, enhancing their livelihood and job satisfaction.

Carryt’s operations, including over 200,000 deliveries per month in 2021, are projected to scale to 1 million deliveries per month in 2022, showcasing the project’s transformative potential.

Future Implications

This Route Optimization Project demonstrates the power of AI and open collaboration in addressing real-world logistics challenges. Future directions include:

  • Enhanced Data Quality: Developing better datasets with fewer restrictions to improve the accuracy and scalability of optimization algorithms.
  • Real-Time Optimization: Integrating dynamic updates for real-time traffic and weather conditions to further improve route efficiency.
  • Broader Application: Expanding the tool’s implementation across Latin America and beyond to support growing e-commerce demands sustainably.

The project exemplifies how technology and teamwork can drive innovative solutions for urban logistics, benefiting businesses, drivers, and the environment alike.

This challenge has been hosted with our friends at
Carryt


Plant Nursery
Monitoring Plants Health with AI and Computer Vision
Shopping for fabric, Lomé, Togo. Photo by Brittany Danisch
Building an AI-powered System to Enhance Economic Policymaking With a Pan BBC African Think Tank
3 women standing and 1 woman sitting in a wheelchair in front of many flags from different countries. These women are part of Fight For Right - a DLO that WID assisted in crisis response.
AI-Driven Resource Identification and Matching System

Become an Omdena Collaborator

media card
Visit the Omdena Collaborator Dashboard Learn More