PPlanner

PPlanner is a classical planner implementing parallel heuristic search algorithms.

Reference

Kuroiwa R, Fukunaga A. 2020. Analyzing and Avoiding Pathological Behavior of Parallel Best-First Search. Proc. the 30th International Conference on Automated Planning and Scheduling. (ICAPS 2020)

Build

Docker

The docker container does not support MPI based distributed parallel search methods.

docker build . -t pplanner

Singularity

sudo singularity build pplanner.sif Singularity

Manual Build

Please see Dockerfile and Singularity to know the commands to build.

Examples

Input a SAS+ file and a config file. The planner generates sas_plan after finding a solution.

GBFS using the FF heuristic.

./bin/planner -f output.sas -c config_examples/GBFS-FF.json

HDGBFS using the FF heuristic.

mpirun -n 16 ./bin/mpi_planner -f output.sas -c config_examples/HDGBFS-FF.json

NE using the FF heuristic.

mpirun -n 16 ./bin/mpi_planner -f output.sas -c config_examples/NE-FF.json

LG using the FF heuristic.

mpirun -n 16 ./bin/mpi_planner -f output.sas -c config_examples/LG-FF.json

KPGBFS using the FF heuristic.

./bin/planner -f output.sas -c config_examples/KPGBFS-FF.json

P_{GBFS} using the FF heuristic.

./bin/planner -f output.sas -c config_examples/PGBFS-FF.json

P_{GBFS}/C using the FF heuristic.

./bin/planner -f output.sas -c config_examples/PGBFS-C-FF.json

PUHF using the FF heuristic.

./bin/planner -f output.sas -c config_examples/PUHF-FF.json

SPUHF using the FF heuristic.

./bin/planner -f output.sas -c config_examples/SPUHF-FF.json