The 2013 competition

Entering the competition

The competition is now open. Read instructions, and register your team to participate.

The competition will end after June 3rd.

ICML Workshop

The competition is associated with an ICML Workshop, where the winners of each category will have the opportunity to present their algorithms. In addition, we welcome submissions regarding empirical evaluation of RL algorithms. The workshop will also discuss the future of the competition.


To suggest or submit domains, offer to help with the organisation, donate a prize, or just general comments on the competition, please email with the subject containing the string "RL Competition 2013".


After a four year hiatus, the reinforcement-learning competition is back. The primary aim of the competition is to test both general and domain-specific reinforcement learning algorithms, using an unbiased and transparent methodology. As a side-effect, the competition will generate a set of benchmarks domains and benchmark results in those domains, which can then be used as a basis of comparison in future work.

As in the previous years, the competition will be hosted at 

The format will be remain as is. Agents may compete in one or more of a set of known domains. This set will include a polyathlon event, where the agents compete in a sequence of arbitrary environments.

We invite: 
  1. Arnaud Barré - server support
  2. Christos Dimitrakakis - overall organisation
  3. Guangliang Li - website and publicity
  4. Nikolaos Tziortziotis - general domain support