The Theoretical Foundation for Incremental Least-Squares Temporal Difference Learning

• Author(s) / Creator(s)

  • Zinkevich, Martin

• Technical report TR06-25. In this paper we present a mathematical foundation for Incremental Least-Squares Temporal Difference Learning (iLSTD) for policy evaluation in reinforcement learning with linear function approximation. iLSTD is an incremental method for achieving results similar to LSTD, the data-efficient, least-squares version of temporal difference learning, without incurring the full cost of the LSTD computation. Here, we give a technical foundation for the asymptotic properties of iLSTD. | TRID-ID TR06-25

• Date created

  2006

• Subjects / Keywords

  • Incremental Least-Squares Temporal Difference Learning

• Type of Item

  Report

• DOI

  https://doi.org/10.7939/R3PR7MW1W

• License

  Attribution 3.0 International