Neuroevolution strategies for episodic reinforcement learning

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@Article{HeidrichMeisner2009152,

• author = "Verena Heidrich-Meisner and Christian Igel",
• title = "Neuroevolution strategies for episodic reinforcement learning",
• journal = "Journal of Algorithms",
• year = "2009",
• volume = "64",
• number = "4",
• pages = "152--168",
• month = oct,
• note = "Special Issue: Reinforcement Learning",
• keywords = "genetic algorithms, genetic programming, Reinforcement learning, Evolution strategy, Covariance matrix adaptation, Partially observable Markov decision process, Direct policy search",
• ISSN = "0196-6774",
• broken = "http://www.sciencedirect.com/science/article/B6WH3-4W7RY8J-3/2/22f7075bc25dab10a8ff3714e2fee303",
• DOI = "doi:10.1016/j.jalgor.2009.04.002",
• abstract = "Because of their convincing performance, there is a growing interest in using evolutionary algorithms for reinforcement learning. We propose learning of neural network policies by the covariance matrix adaptation evolution strategy (CMA-ES), a randomised variable-metric search algorithm for continuous optimisation. We argue that this approach, which we refer to as CMA Neuroevolution Strategy (CMA-NeuroES), is ideally suited for reinforcement learning, in particular because it is based on ranking policies (and therefore robust against noise), efficiently detects correlations between parameters, and infers a search direction from scalar reinforcement signals. We evaluate the CMA-NeuroES on five different (Markovian and non-Markovian) variants of the common pole balancing problem. The results are compared to those described in a recent study covering several RL algorithms, and the CMA-NeuroES shows the overall best performance.",
• notes = "compared against CE \cite{gruau:1996:ceVdeGNN}",

}

Genetic Programming entries for Verena Heidrich-Meisner Christian Igel

Citations