HQ-LEARNING

Authors
WIERING M SCHMIDHUBER J
Citation
M. Wiering et J. Schmidhuber, HQ-LEARNING, Adaptive behavior, 6(2), 1997, pp. 219-246
Citations number
53
Journal title
Adaptive behavior → ACNP
ISSN journal
10597123
Volume
6
Issue
2
Year of publication
1997
Pages
219 - 246
Database
ISI
SICI code
1059-7123(1997)6:2<219:>2.0.ZU;2-P
Abstract
HQ-learning is a hierarchical extension of Q(lambda)-learning designed to solve certain types of partially observable Markov decision proble ms (POMDPs). HQ automatically decomposes POMDPs into sequences of simp ler subtasks that can be solved by memoryless policies learnable by re active subagents. HQ can solve partially observable mazes with more st ates than those used in moss previous POMDP work.