This is a repository for the following paper:
- Kazumi Kasaura. 2023. “Homotopy-Aware Multi-Agent Path Planning in Plane.”
You can use Dockerfile to build an environment.
mkdir build
cd build
cmake ..
make
The following commands are to be run on the build folder.
mkdir grid_instances
bash ../script/generate_grid_instances.sh 29 29 40 10
mkdir logs
bash ../script/record_logs.sh 29 29 40 10 100
python3 ../script/show_runtimes.py
python3 ../script/show_maxcd.py
mkdir cd_values
./dynnikov_count 3 30 100 100 > cd_values/3.txt
./dynnikov_count 5 50 100 100 > cd_values/5.txt
./dynnikov_count 10 100 100 100 > cd_values/10.txt
./dynnikov_count 100 1000 100 100 > cd_values/100.txt
./dynnikov_count 1000 10000 100 100 > cd_values/1000.txt
python3 ../script/show_cd_values.py
bash ../script/generate_grid_instances.sh 14 14 10 100
mkdir PP_results
bash ../script/solve_by_PP.sh 14 14 10 100 100
mkdir PPvP_results
bash ../script/solve_by_PPvP.sh 14 14 10 100 100
Although this repository does not contain any optimal MAPF solver for the Optimal One baseline,
results of such a solver must be stored in a folder named CBS_results
for the following steps.
mkdir PP_opt_results
mkdir PPvP_opt_results
mkdir CBS_opt_results
bash ../script/optimize_PP_and_CBS.sh 14 14 10 100 100
python3 ../script/show_results.py
python3 ../script/count_braids.py > braid_counts.txt
python3 ../script/show_histogram.py
This software is released under the MIT License, see LICENSE.
@misc{kasaura2024homotopyaware,
title={Homotopy-Aware Multi-Agent Path Planning in Plane},
author={Kazumi Kasaura},
year={2024},
eprint={2310.01945},
archivePrefix={arXiv},
primaryClass={cs.MA}
}