Multi-Agent Pickup and Delivery
Implementation of the TA-Prioritized algorithm with a modified deadlock avoidance method for a significant improvement of the larger instances' makespans. This implementation does not compute the solution for the travelling salesman problem, but instead loads the solution provided by the author.
Orange cells: Agent parking locations
Blue cells: Task endpoints (pickup/delivery)
Black cells: Wall/blocked cell
Requires Mesa to run the simulation.
Install with PIP:
pip install mesa
Install with Conda:
conda install mesa