Skip to content

Quick start#

This page provides you with different options for setting up Fast Downward on your machine.

Fast Downward is released in four flavours: Tarball, Apptainer (formerly known as Singularity), Docker and Vagrant. Here we provide instructions to get you started as quickly as possible. You can find more usage information at planner usage.

What flavour is for me?#

  • running experiments: We recommend Apptainer or the tarball. Docker is an alternative, but be mindful of its significant overhead.
  • teaching: We recommend Vagrant.
  • development: We recommend working on a clone of the master repository.

See what-flavour-is-for-me for a more detailed discussion.

Apptainer#

Apptainer Installation#

To run the Apptainer image, first install one of the Apptainer releases or follow the steps provided on the Apptainer page. We tested with versions 1.2.2 and 1.4.1.

To download the Fast Downward image, run:

apptainer pull fast-downward.sif docker://aibasel/downward:latest

Apptainer Usage#

You can run the planner as follows:

./fast-downward.sif <your-planner-options>

Assume you want to solve a planning task (e.g. from the Fast Downward benchmarks) using A* search with the LM-cut heuristic. To do so you can run the following:

./fast-downward.sif <domain.pddl> <task.pddl> --search "astar(lmcut())"

See more planner options at planner usage.

Docker#

Docker Installation#

To install Docker on your machine follow the steps provided on the Docker page.

Docker Usage#

You can run the planner as follows:

sudo docker run aibasel/downward <your-planner-options>

Note:

  • The use of sudo (Docker usually requires root privileges).
  • The Docker image for Fast Downward is installed on your machine as a side-effect of the command.
  • You can use the Docker flag -rm for cleaning up the container (recommended)

Assume you want to solve a planning task (e.g. from the Fast Downward benchmarks) using A* search with the LM-cut heuristic. To do so you can run the following:

sudo docker run --rm -v <path-to-benchmarks>:/benchmarks aibasel/downward \
/benchmarks/<domain.pddl> /benchmarks/<task.pddl> \
--search "astar(lmcut())"

Here <path-to-benchmarks> refers to your local directory containing your domain and task file.

Note:

  • The Docker flag -v mounts the local directory <path-to-benchmarks> of your host machine under the container directory /benchmarks, which is the place where the containerised planner looks for the problem.
  • <path-to-benchmarks> must be an absolute path; relative paths are only supported as of Docker Engine version 23.

See more planner options at planner usage.

Vagrant#

Vagrant Installation#

To install Vagrant on your machine follow the steps provided on the Vagrant page. For Vagrant to work you also need a "provider" such as VirtualBox, VMware, Hyper-V.

Note:

  • Not all provider versions and Vagrant versions are compatible!
  • We recommend using Vagrant 2.4.1 and VirtualBox 7.0.20.

To create your virtual machine using Vagrant create a new directory <my-fast-downward-vm> containing only the Fast Downward Vagrantfile for the desired release.

If you want to use LP solvers follow the steps under LP Solvers.

Otherwise, you can set up your virtual machine directly from the <my-fast-downward-vm> directory using:

vagrant up

LP solvers#

The SoPlex LP solver is included automatically. If you want to also use the CPLEX LP solver within the planner, its installer file must be present in the directory <path-to-lp-installers>. As of Fast Downward 23.06, you will need CPLEX 22.1.1 (installer filename cplex_studio2211.linux-x86-64.bin).

To set up your virtual machine from the <my-fast-downward-vm> directory as follows:

export DOWNWARD_LP_INSTALLERS=<path-to-lp-installers>

vagrant up

You can now safely delete the LP installers and unset the environment variable.

Vagrant Usage#

You can run the planner from within the vagrant VM as follows:

downward/fast-downward.py <your-planner-options>

Assume you want to solve a planning task (e.g. from the Fast Downward benchmarks) using A* search with the LM-cut heuristic. To do so you can run the following:

# From the <my-fast-downward-vm> directory 

# Upload the benchmarks to the benchmarks directory on your vagrant machine
vagrant upload <path-to-benchmarks> benchmarks/

# Log into the VM.
vagrant ssh

# Run the planner.
downward/fast-downward.py \
/vagrant/benchmarks/<domain.pddl> /vagrant/benchmarks/<task.pddl> \
--search "astar(lmcut())"

# Log out from the VM.
logout

# Shut down the machine
vagrant halt

Here <path-to-benchmarks> refers to your local directory containing your domain and task file.

See more planner options at planner usage.

Source Code#

Source Code Installation#

We recommend using the latest release, especially for scientific experiments. See the build instructions for a complete description on how to build the planner from source.

Tarball#

Download the tarball from the latest release. E.g. to build Fast Downward 24.06.01 download the tarball from the Fast Downward 24.06 release page and run the following: 

tar -xvzf fast-downward-24.06.1.tar.gz
cd fast-downward-24.06.1
./build.py

Repository#

Alternatively, you can fork/clone the repository. If you are using the main branch, be aware that things can break or degrade with every commit.

Source Code Usage#

You can run the planner as follows:

./fast-downward.py <your-planner-options>

Assume you want to solve a planning task (e.g. from the Fast Downward benchmarks) using A* search with the LM-cut heuristic. To do so you can run the following:

./fast-downward.py <domain.pddl> <task.pddl> --search "astar(lmcut())"

See more planner options at planner usage.

Next steps#