GOOSE: Graphs Optimised fOr Search Evaluation

GOOSE is a learning for planning framework. It contains various methods for learning representations for planning tasks, and algorithms for using such representations for solving planning tasks.

• If you just want to use the WL features, take a look at the WLPlan package. The dev branch illustrates how to integrate it into both learning and planning code.
• If you want the code used for our NeurIPS-24 paper, refer to the NeurIPS-24 release.
• If you want the code used for our ICAPS-24 paper, refer to the ICAPS-24 release.
• If you want the code used for our AAAI-24 paper, refer to the AAAI-24 release.
• Future releases will deprecate support for GNN models and will be built around the WLPlan package.

See references for the corresponding publications.

Table of contents

• GOOSE: Graphs Optimised fOr Search Evaluation
  • Table of contents
  • Setup
  • Training
    • Example for WL models
    • Example for GNN models
  • Planning
    • Example for WL models
    • Example for GNN models
  • References

Setup

Use the commands below to make a virtual environment, activate it, install packages, and build cpp components. The setup has been tested with python versions 3.10 and higher, but should probably work for lower python3 versions as well.

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
sh setup.sh

In case a virtual environment does not work, you can also try anaconda:

conda create --name goose python=3.10.4
conda activate goose
pip install -r requirements.txt
sh setup.sh

Training

• see python3 train.py -h for help, you will need the --save-file argument if you want to save the model
• to train with your own dataset, you will need to construct an experiment configuration toml file such as in here
  • the tasks_dir and plans_dir paths must contain the same files, differentiating only in the file suffix (.pddl and .plan, respectively)

Example for WL models

python3 train.py experiments/models/wl_ilg_gpr.toml experiments/ipc23-learning/blocksworld.toml --save-file blocksworld_wl.model

Example for GNN models

python3 train.py experiments/models/gnn_mean_ilg.toml experiments/ipc23-learning/blocksworld.toml --save-file blocksworld_gnn.model

Planning

• see run_wl.py for WL models and run_gnn.py for GNN models
• GNN models automatically try to use GPU where possible and CPU otherwise

Example for WL models

python3 run_wl.py benchmarks/ipc23-learning/blocksworld/domain.pddl benchmarks/ipc23-learning/blocksworld/testing/medium/p01.pddl blocksworld_wl.model

Example for GNN models

python3 run_gnn.py benchmarks/ipc23-learning/blocksworld/domain.pddl benchmarks/ipc23-learning/blocksworld/testing/medium/p01.pddl blocksworld_gnn.model

References

The relevant publications for this repository are:

• Dillon Ze Chen and Sylvie Thiébaux. Learning Heuristics for Numeric Planning. NeurIPS 2024. [coming soon]
• Dillon Ze Chen and Felipe Trevizan and Sylvie Thiébaux. Return to Tradition: Learning Reliable Heuristics with Classical Machine Learning. ICAPS 2024. [pdf | bib]
• Dillon Ze Chen and Sylvie Thiébaux and Felipe Trevizan. Learning Domain-Independent Heuristics for Grounded and Lifted Planning. AAAI 2024. [pdf | bib]