Hackathon Mit Mcp Or

What is Hackathon Mit Mcp Or

hackathon_MIT_MCP-OR is an interactive application developed during the Global AI MIT Hackathon, designed to compute the optimal freight transport route between two cities based on user-defined priorities such as speed, cost, and environmental impact.

Use cases

Use cases include optimizing delivery routes for logistics companies, planning freight transport for events, and assessing the environmental impact of different transportation methods.

How to use

To use hackathon_MIT_MCP-OR, clone the repository, install the necessary dependencies, create and sync a virtual environment, and run the backend server using ‘uvicorn’. Then, start the frontend interface with npm.

Key features

Key features include the ability to select any two cities, adjust preferences for time, price, and sustainability, support for multi-modal transportation (road, air, sea), utilization of real-world data for route planning, and dynamic updates based on live conditions.

Where to use

hackathon_MIT_MCP-OR can be used in logistics, transportation planning, environmental studies, and any industry that requires efficient freight routing.

Global AI MIT Hackathon Project: Optimal Freight Route Planner
Hackathon entry for the Global AI MIT JHackathon

Team Members (École Polytechnique)

• Mohamed Aloulou [email protected]
• Hippolyte Wallaert [email protected]
• Antoine Maecheler [email protected]
• Édouard Rabasse [email protected]

Project Overview

We have built an interactive application that computes the optimal freight transport route between two cities based on user-definable priorities: speed, cost, and environmental impact (CO₂ emissions).

• Origins & Destinations: Users select any two cities.
• Preferences: Adjust sliders for time, price, and sustainability importance.
• Multi-Modal Network: Supports road, air, and sea segments.
• Real-World Data: Leverages geographical coordinates, transportation timetables, CO₂ emission estimates, and price data.
• Dynamic Updates: Incorporates live traffic, weather, or other user-provided conditions to recalculate the best route in real time.

Run the interface

Getting started

Clone the repository and install dependencies:

1. Navigate to the project directory:

   cd <repo_name>
   

2. This project depends on the uv CLI. Please ensure it’s installed—if not, install it as follows :

• MacOS (Homebrew):

brew install uv

• Debian/Ubuntu:

# For Debian/Ubuntu-based systems
sudo apt update && sudo apt install uv

• Windows (Chocolatey):

choco install uv

3. Create & Sync Virtual Environment

   uv sync
   

Running the interface

Run the backend server

   uv run uvicorn main:app --reload

Then in a new terminal, run the frontend:

    npm start

Note : Ensure Node.js and npm are installed. Download them from Node.js official website.

Use Claude via our custom MCP server

We added Claude for real-time change detection and to provide the user with the best route at any time. The user can input traffic information, weather conditions, and other factors that may affect the route.

This functionality currently only works on macOS. Please make sure you’ve downloaded the Claude Desktop app (https://claude.ai/) and have an active subscription. Then install and run the MCP server via:

uv run mcp install csv_editor.py --name "Shipping MCP" --with-editable .

Project Structure

Backend

The first part of the project was gathering data and creating a database of cities and routes.
Initially, we picked 30 US cities and used open source APIs to gather data on their geographical coordinates, airports, and ports. We then generated a list of all possible routes between these cities, including road, air, and ship routes. The data was enriched with estimated travel times, distances, CO2 emissions, and prices.
This grid allows us to cover the US territory but should the user input a city that is not in the list, the program will add it to the database and generate the routes for it.

Solver

The second part was translating the problem into a linear programming problem. We used gurobipy to create a model that minimizes the cost of transportation while respecting the constraints of time and CO2 emissions. The model takes into account the user’s preferences for speed, price, and sustainability.
The solver uses the data from the database to find the optimal route between the two selected cities.

Then we compute the best alternative routes using the same model. The user can select the best route based on their preferences, and the program will provide the best alternative routes as well.

Frontend

Then we developped a tool to visualize the results. The user can select the origin and destination cities, adjust the importance of time, cost, and CO2 emissions, and see the best route on a map.