This is the code implementing symbolic skill set extension for an agent relying on symbolic planning.
There are two options to run the code: either natively or in a docker container (no GUI available).
For this, please follow the instructions in highlevel_planning_ros/README.md.
For this, a copy of Docker Desktop needs to be installed on your machine (https://www.docker.com/get-started).
Download version 2.1 of the Metric-FF symbolic planner from https://fai.cs.uni-saarland.de/hoffmann/metric-ff.html, extract it, and place it next to this file, such that the directory high_level_planning/Metric-FF-v2.1 contains the Metric-FF source code.
Build the Docker container:
docker build -t hlp_docker .Launch the container:
docker run -it --rm hlp_dockerYou should end up in a bash shell within the container. From here, our exploration algorithm can be run (only in direct mode since no GUI is available from within the container):
| Domain | Algorithm | Command |
|---|---|---|
| PDDL benchmark domain | ours | python3 scripts/run_pddl_benchmark_ours.py |
| PDDL benchmark domain | MCTS | python3 scripts/run_pddl_benchmark_mcts.py |
| Rearrangement task domain | ours | python3 scripts/run_auto.py -m direct |
| Rearrangement task domain | MCTS | python3 scripts/run_mcts.py -m direct |
More details on changing configurations, etc. can be found in the other readme.