291 search results - page 4 / 59 » Implementation with a bounded action space |
WDAG
2007
Springer
14 years 10 days ago
2007
Springer
The timestamp problem captures a fundamental aspect of asynchronous distributed computing. It allows processes to label events throughout the system with timestamps that provide in...
JAIR
2008
13 years 6 months ago
2008
Probabilistic planning problems are typically modeled as a Markov Decision Process (MDP). MDPs, while an otherwise expressive model, allow only for sequential, non-durative action...
AAAI
1997
13 years 7 months ago
1997
Partially observable Markov decision processes (POMDPs) allow one to model complex dynamic decision or control problems that include both action outcome uncertainty and imperfect ...
Publication
Abstract: Several approximate policy iteration schemes without value functions, which focus on policy representation using classifiers and address policy learning as a supervis...
GECCO
2005
Springer
13 years 11 months ago
2005
Springer
An experimental form of Modulation (Reinterpretation) of the Search Space is presented. This modulation is developed as a mathematical method that can be implemented directly into...