mushfiqur11 / intersectionmanagementai

The AI model was built to optimize traffic at any 4-road intersection (6 lanes each)

Automated Intersection Management - Heuristic Approach (with MiniZinc)

[paper] [code]

• Md. Mushfiqur Rahman
• Nahian Muhtasim Zahin
• Kazi Raiyan Mahmud
• Md. Azmaeen Bin Ansar

Dependencies

• MiniZinc
• Python Notebook
• iminizinc library (for integrating minizinc with python)
• drawSVG (for visualizing data)

Description

The IntersectionManagement.mzn file contains the model of the AI. The file takes in an array of old paths named "oldPaths" and an array with the old queue conditions named "WaitingQueue". As output, the model creates a new "paths" array and a "NewWaitingQueue" array. These two arrays are used for the next iteration. The python notebook doesn't work yet. To run the system, the model needs to be manually called repeatedly.

Instruction

Install MiniZinc

For linux-based OS please use the following code on terminal

sudo apt-get update
sudo apt-get install minizinc

For Mac

• Launch Terminal by pressing command+space, type terminal and hit Enter key.
• Run

ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)" < /dev/null 2> /dev/null

• Run

brew install minizinc

For Windows The python and minizinc integration is still at an early stage. And windows support is still not properly available yet.

Install Python Notebook

Use any python notebook. Jupyter notebook can be installed through anaconda[https://www.anaconda.com/].

Install iminizinc library

Run the following code to install iminizinc[https://github.com/MiniZinc/iminizinc] extension

pip install -U iminizinc

Make sure that the mzn2fzn binary and the solvers are at the PATH when running iminizinc on python notebook.

Install drawSVG

Run the following code in terminal to install drawSVG[https://pypi.org/project/drawSvg/]

pip install drawSVG