Skip to content

hejm37/sysu-planner

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

sysu-planner

The SYSU-Planner is a two-stage planner designed to solve classical planning problems. It first performs the 1-BFWS (Nir and Hector 2017) with very fast speed. If it fails to find a solution, it will then perform a modified online refinement algorithm named Forward-RHC (see also Maximilian and Jorg 2018).

Build and run with container

Using the planner with Singularity is rather simple. First install Singularity following this guide. Then run the following script in CLI and you will have the plan file sas_plan under $RUNDIR.

sudo singularity build planner.img sysu-planner/Singularity
mkdir rundir
cp path/to/domain.pddl rundir
cp path/to/problem.pddl rundir
RUNDIR="$(pwd)/rundir"
DOMAIN="$RUNDIR/domain.pddl"
PROBLEM="$RUNDIR/problem.pddl"
PLANFILE="$RUNDIR/sas_plan"
singularity run -C -H $RUNDIR planner.img $DOMAIN $PROBLEM $PLANFILE $COSTBOUND

Supported problems

The formulation of supported problems is the same as IPC 2018. We also provide a set of supported domains and problems in benchmark-domains.

Notes on playing with the source code

The source code of the planner contains two part:

  • BFWS-public and its dependency, LAPKT-public
  • fast-downward-conjunctions

Then planner should be invoked in the fast-downward-conjunctions part (using --dual option and it will call BFWS-public/fd-version/bfws.py to perform 1-BFWS, see the Singularity script for more details).

Potential Failures

If the above build has failed, it may appears to be a cmake cache fail. In this case, remove the builds (if it exists) directory under fast-downward-conjunctions and rerun the singularity command shall solve the problem.

Or it may appears to be a scons build fail. In this case, remove all the .sconsign.dblite files under the directory shall solve the problem.

Both cases would occur if the planner was built outside a container.

About

sparkle planning challenge

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published