PDDL4J library

1. Contact

• Damien Pellier ([email protected])
• http://lig-membres.imag.fr/pellier/

2. Description

PDDL4J is an open source library under LGPL license.

The purpose of PDDL4J is to facilitate the development of JAVA tools for Automated Planning based on PDDL language (Planning Domain Description Language). Automated planning and scheduling, in the relevant literature often denoted as simply planning, is a branch of artificial intelligence that concerns the realization of strategies or action sequences, typically for execution by intelligent agents, autonomous robots and unmanned vehicles.

PDDL was originally developed by Drew McDermott and the 1998 planning competition committee. It was inspired by the need to encourage the empirical comparison of planning systems and the exchange of planning benchmarks within the community. Its development improved the communication of research results and triggered an explosion in performance, expressivity and robustness of planning systems.

PDDL has become a de facto standard language for describing planning domains, not only for the competition but more widely, as it offers an opportunity to carry out empirical evaluation of planning systems on a growing collection of generally adopted standard benchmark domains. The emergence of a language standard will have an impact on the entire field, influencing what is seen as central and what peripheral in the development of planning systems.

The library contains:

• A PDDL 3.1 parser and all the classes need to manipulate its concepts. The parser can be configured to accept only specified requirements of PDDL language.
• A set of useful pre-processing mechanisms in order to instantiate and simply operators into ground actions based on inertia properties.
• A set of already implemented classical heuristics, e.g., h_ff, h_max, etc.
• Several examples of planners using PDDL4J.

3. How to use the PDDL4J library?

The pddl4j library is package with "gradle". If ant is not installed on your computer, go to https://gradle.org/ and install it.

3.1 How to build PDDL4J?

Type at the root of the PDDL4J distribution:

./gradlew build

if you want to build the project with findbug:

./gradlew build -Pfindbug

The jar "pddl4j-VERSION.jar" is available in the build/libs directory.

3.3 How to use binary distribution?

The pddl4j library can be used as a jar file called "pddl4j-VERSION.jar" in the "libs" directory of the release. Therefore, include the jar in our classpath and enjoy it.

3.4 How to access the documentation?

All the documentations are contained in the doc directory.

To generate the javadoc just type:

./gradlew javadoc

3.5 How to run existing planner?

Planners are available in the "planners" package of the distribution. For instance, this archive contains a simple planner based on A* search strategy called HSP. To launch this planner use the following command line:

java -javaagent:build/libs/pddl4j-3.7.0.jar -server -Xms2048m -Xmx2048m fr.uga.pddl4j.planners.statespace.StateSpacePlannerFactory -o pddl/blocksworld/domain.pddl -f pddl/blocksworld/p15.pddl

java -jar build/libs/pddl4j-3.7.0.jar -o pddl/blocksworld/domain.pddl -f pddl/blocksworld/p15.pddl

Or use the gradle run command:

gradle run -PArgs=-o,pddl/blocksworld/domain.pddl,-f,pddl/blocksworld/p15.pddl

Note: A set of planning problems is available in the web site of the international planning competition: http://ipc.icaps-conference.org.

4. How to cite

PDDL4J: a planning domain description library for java https://doi.org/10.1080/0952813X.2017.1409278

D. Pellier & H. Fiorino (2017) PDDL4J: a planning domain description library for java, Journal of Experimental & Theoretical Artificial Intelligence, 30:1, 143-176, DOI: 10.1080/0952813X.2017.1409278

5. Changelog

PDDL4J v3.7.0

Update project tools

• log4j 2.11
• Checkstyle 8.9
• SonarQube 2.6.1
• Javacc 2.4
• Gradle wrapper 4.8
• Add Jenkinsfile script for CI

Planner

• New package: Statespace with HSP and FF planners
• Add StateSpacePlanner and StateSpacePlannerFactory interfaces
• Use StateSpacePlannerFactory to create state space planners

State space strategy

• Search strategies are now independant from planners: Add StateSpaceStrategy interface
• A*, Greedy Best First Search, Enforced Hill Climbing and Hill Climbing are available
• Add JUnit tests

Global

• Add tests
• Fix bugs and javadoc
• Memory Agent: PDDL4J could work even if JVM command line is not given
• Big work on JUnit tests and VAL

PDDL4J v3.6.0

Planner

• Adding Fast Forward planner
• Adding Enforced Hill Climbing planner
• Adding depth attibute in Node class
• Converting cost and heuristic in double
• Fix bugs

Parser

• Adding parseFromString method (it's is now possible to parse a string containing a pddl domain and/or problem)
• Refactor Parser class and fix bugs

JUnit tests

• Adding tests for all the planners, CodedProblem, Encoder, Parser, Plan, etc.
• Refactor and fix bugs

Global

• Change AdaptatorJavaJson to JsonAdapter
• Fix logger and standard outputs
• Throws uncatched exceptions
• Fix bugs