Bellman goes Relational

Kristian Kersting, M. van Otterlo, Luc De Raedt

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    71 Citations (Scopus)
    159 Downloads (Pure)


    Motivated by the interest in relational reinforcement learning, we introduce a novel relational Bellman update operator called ReBel. It employs a constraint logic programming language to compactly represent Markov decision processes over relational domains. Using ReBel, a novel value iteration algorithm is developed in which abstraction (over states and actions) plays a major role. This framework provides new insights into relational reinforcement learning. Convergence results as well as experiments are presented.
    Original languageEnglish
    Title of host publicationProceedings of the International Conference on Machine Learning (ICML'04)
    EditorsR. Greiner, D. Schuurmans
    Place of PublicationNew York
    PublisherUniversity of Alberta
    Number of pages8
    ISBN (Print)1-58113-8385
    Publication statusPublished - 7 Dec 2004
    Event21st International Conference on Machine Learning, ICML 2004 - Banff, Canada
    Duration: 4 Jul 20048 Jul 2004
    Conference number: 21


    Conference21st International Conference on Machine Learning, ICML 2004
    Abbreviated titleICML


    • EC Grant Agreement nr.: FP6/508861
    • HMI-IA: Intelligent Agents


    Dive into the research topics of 'Bellman goes Relational'. Together they form a unique fingerprint.

    Cite this