118
Voted
CORR
1998
Springer
15 years 2 months ago
1998
Springer
A new training algorithm is presented for delayed reinforcement learning problems that does not assume the existence of a critic model and employs the polytope optimization algorit...