126
click to vote
ICML
2004
IEEE
16 years 3 months ago
2004
IEEE
Synchronous reinforcement learning (RL) algorithms with linear function approximation are representable as inhomogeneous matrix iterations of a special form (Schoknecht & Merk...