reinforcement learning test
multi_astar.py and single_path.py read their obstacle grids from a plain
text file. This repository includes a default map.txt that both scripts load
automatically. To use a different map, set the MAP_FILE environment variable
to another file path. Each map file should contain rows of 0 and 1
separated by spaces, where each line corresponds to one row in the grid. For
example:
1 1 1 1
1 0 0 1
1 0 0 1
1 1 1 1
If the specified file cannot be read, the scripts raise a clear error message.
single_path.py samples random start/goal pairs from the map and runs A* for
each agent. The number of agents defaults to six but can be changed with the
--agents command-line option. When executed it opens a Matplotlib window and
animates all agents moving along their computed paths, each agent shown in a
different color with a numbered label inside a larger marker.