NumericTCORE

Repository Structure

• bin/: the compiler's executable
• benchmark/: benchmarks instances
• NumericTCORE/: implementation of Numeric Tcore

Installation

Numeric Tcore uses the following external libraries:

• unified-planning version 1.1.0
• click
• sympy

Use the following command to install Numeric Tcore. We suggest starting with a fresh Python environment (we tested with Python 3.10).

From the "NumericTCORE" directory, install the package with:

pip install . --no-cache-dir

The additional libraries will be installed automatically.

Running the compilation

The main entry point of Numeric Tcore is "./bin/ntcore.py". Basic usage:

python ./bin/ntcore.py DOMAIN PROBLEM OUTPUT_PATH

• DOMAIN is the path to the pddl domain file.
• PROBLEM is the path to the pddl problem file.
• OUTPUT is the path to the folder (which must exist) that will contain the compiled domain and problem files.

To switch between the two variants of Numeric Tcore, use one of the following commands:

python ./bin/ntcore.py DOMAIN PROBLEM OUTPUT --naive_mode
python ./bin/ntcore.py DOMAIN PROBLEM OUTPUT --delta_mode

The option --naive_mode runs the basic version of ntcore, while --delta_mode runs ntcore^+. The default is --naive_mode. For more informations, run python ./bin/ntcore.py --help

Limitations

The current version of Numeric Tcore has been thoroughly tested with trajectory constraints featuring formulas with only numeric conditions (those in the benchmark/). In principle, Numeric Tcore supports propositional constraints, but in practice, such a feature has only been briefly tested.

Example

While inside the NumericTCORE folder, to compile the first instance of counters with ntcore^+ you can run:

python ./bin/ntcore.py benchmark/fn-counters/domain.pddl benchmark/fn-counters/instance_2.pddl . --delta_mode

Citation

@inproceedings{DBLP:conf/aaai/BonassiGS24,
  author       = {Luigi Bonassi and
                  Alfonso Emilio Gerevini and
                  Enrico Scala},
  title        = {Dealing with Numeric and Metric Time Constraints in {PDDL3} via Compilation
                  to Numeric Planning},
  booktitle    = {{AAAI}},
  pages        = {20036--20043},
  publisher    = {{AAAI} Press},
  year         = {2024}
}