Skip to content

danielzgsilva/MultiAgentPathFinding

Repository files navigation

Multi Agent Path Finding Visualization

This is a visualization of multi-agent path finding on 3D terrain maps

An example visualization

Usage

This visualization runs on Python 3.7

Start off by cloning the repo:
git clone https://github.com/danielzgsilva/MultiAgentPathFinding

Navigate to the project's root and install dependencies into a conda environment like so:
conda env create -f environment.yml

Activate the environment
conda activate path_finding

Run the visualization with
python run.py

Available commands are:

  • --num_agents | number of agents to spawn into the world | REQUIRED
  • --starts | list of x, y start coordinates for each agent | REQUIRED
  • --goals | list of x, y goal coordinates for each agent | REQUIRED
  • --max_time | maximum amount of time steps each agent has to each its goal | REQUIRED
  • --width | width of the grid
  • --height | height of the grid
  • --map_type | type of map to create (plains, mountains, canyons, default - combined)
  • --no_viz | if set, the visualization will not run, only a list of paths will be returned

Full run statement like so:
python run.py --num_agents 2 --starts 0,0 5,5 --goals 49,49 45,45 --max_time 30

Work in progress:

  • Creating PyQt GUI
  • Cooperative A* algorithm runtime improvements
  • Scheduling agents between multiple goals
  • Implementing Windowed Hierarchical Cooperative A*

About

Multi-agent cooperative A* path finding visualization using PyQtGraph and OpenGL

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages