Source code used in the experiments from the paper "Grounding Planning Tasks Using Tree Decompositions and Iterated Solving" (Corrêa et al ICAPS 2023). There are six different files: asp-grounding.zip: It contains the files to ground planning tasks using off-the-shelf ASP tools. The README.md file in this zip explains how to execute it. (You also need to add lpopt to the PATH and to an environment variable LPOPT_BIN_PATH to use all options. See lpopt.zip below.) clingo.zip: It contains the versions of clingo and gringo used for the experiments. Just add the 'cling' and 'gringo' executables to the necessary environment variables. downward-new-grounder.zip: It contains a version of Fast Downward using the new grounder via iterated solving. To use this, you need to set some variables first: (i) gringo should be in the PATH; (ii) LPOPT_BIN_PATH should point to lpopt; (iii) COUNT_GROUND_ACTIONS_BIN_PATH should point to count-ground-actions.py (found in asp-grounding.zip). Then, you can follow the README instruction to build and execute Fast Downward. However, you need to pass the option "--translate-options --use-direct-lp-encoding" to the driver script (i.e., fast-downward.py). Note that this only works with STRIPS. downward-gringo.zip: It contains a version of Fast Downward replacing the Python grounder with gringo. (It is a mistake that some parts of the code refer to 'clingo' instead.) Once again, you need to have gringo in the PATH. Then, you can follow the README instruction to build and execute Fast Downward. However, you need to pass the option "--translate-options --use-direct-lp-encoding" to the driver script (i.e., fast-downward.py). Note that this only works with STRIPS. fd-grounder.zip: It contains a reimplementation of Fast Downward's grounder (Helmert AIJ 2009, section 6) in C++. The README.md file in this zip explains how to execute it. lpopt.zip: It contains the version of lpopt used in our experiments. You need to add to the PATH or set an environment variable pointing to lpopt depending on which other package you want to use it with. As an important remark, we recommend the use of Pypy3 instead of Python 3. (By the time of this submission, the most recent version of Python 3 was 3.11.)