115
Voted
NIPS
2007
15 years 4 months ago
2007
We present an algorithm called Optimistic Linear Programming (OLP) for learning to optimize average reward in an irreducible but otherwise unknown Markov decision process (MDP). O...