Analyzing Feature Generation for Value-Function Approximation
Ronald Parr - Duke University, USA
Christopher Painter-Wakefield - Duke University, USA
Lihong Li - Rutgers University, USA
Michael Littman - Rutgers University, USA
We analyze a simple, Bellman-error-based approach to generating basis functions for valuefunction approximation. We show that it generates orthogonal basis functions that provably tighten approximation error bounds. We also illustrate the use of this approach in the presence of noise on some sample problems.