Probabilistic and Reinforcement Learning Track

2023 International Planning Competition

Welcome to the International Planning Competition 2023: Probabilistic and Reinforcement Learning Track

The International Probabilistic Planning Competition is organized in the context of the International Conference on Planning and Scheduling (ICAPS). It empirically evaluates state-of-the-art planning systems on a number of benchmark problems. The goals of the IPC are to promote planning research, highlight challenges in the planning community and provide new and interesting problems as benchmarks for future research.

Since 2004, probabilistic tracks have been part of the IPC under different names (as the International Probabilistic Planning competition or as part of the uncertainty tracks). After 2004, 2006, 2008, 2011, 2014, and 2018, the 7th IPPC will be held in 2023 and conclude together with ICAPS, in July 2023, in Prague (Czech Republic). This time it is organized by Ayal Taitler and Scott Sanner.


Please forward the following calls to all interested parties:

We invite interested competitors to join the competition discussion:

Preliminary Schedule

Event Date
Infrastructure release with sample domains October, 2022
Call for domains and praticipants October, 2022
Final domains announcement December, 2022
Competitors registration January, 2023
Planner abstract dubmission May, 2023
Contest run June, 2023
Results announced July, 2023


This year’s competition will be using the generic pyRDDLGym - an autogeneration tool for gym environments from RDDL textual description.

RDDLGym diagram

More information about the infrastructure, how to use it and how to add user defined domains can be found the following short guide


There are several original domains at the time of the competition call, but more domains will be added, and we strongly encourage the community to create and contribute used defined domains in you field of expertise, or any problem you think the community should try to solve. The highest quality community contributed domains will be selected to be part of the competition’s domains. We are aware that participants that submit a domain that is used have some advantage with respect to performance on that domain. We view this as a good incentive for teams to submit high quality proposals.

In addition to the original domain, we have recreated some of the classical control domains in RDDL. Illustrating how easy it is to generate domains in pyRDDLSim:

Note, that there are additional domains out there from past competitions (IPPC 2011, IPPC 2014), which can be also be used with pyRDDLSim:

Past compeition were entirely discrete and as the focus of this year is continous and mix discrete-continous problems, their domains were not included in the pyRDDLGym repository. However everybody are welcome to take advantage of their existence. Just make sure they do not make use of something outside of pyRDDLGym RDDL subset. In that case the original Java simulator can be used.


The exact details of the evaluation process are still in discussions (in order to be able to accomodate planing and learning approaches), but the general lines will follow the evaluation of the past IPPC 2011 competition.

Domains (note numbers might change this year)


In addition we intend to make the following changes to the 2011 evaluation (subject to community discussions)

  1. Per-instance time limits instead of an overall 24 hour time limit.
  2. Competitor can do as many simulations as it wants before it initiates the final evaluation of 30 trials.
  3. We’re considering using strong manual baseline policies for max normalization (hence scores will not necessarily be in [0,1])


At this point just join the google group (see link at the bottom) and announce your interest to compete in a post that includes your tentative team name, organization, and team members.



Contact us: