ICML
2004
IEEE
15 years 8 days ago
2004
IEEE
Synchronous reinforcement learning (RL) algorithms with linear function approximation are representable as inhomogeneous matrix iterations of a special form (Schoknecht & Merk...