Automatic Formation of Abstract State Space Representations for Reinforcement Learning Agents Agents with Hierarchical Skill Sets.
Mehran AsadiManfred HuberPublished in: Security and Management (2008)
Keyphrases
- reinforcement learning agents
- state space
- state abstraction
- reinforcement learning
- dynamic environments
- multi agent
- markov decision processes
- multi agent environments
- heuristic search
- reinforcement learning algorithms
- function approximation
- multi agent systems
- cooperative
- markov decision process
- multiple agents
- autonomous agents
- search space
- action selection
- state variables
- transfer learning
- optimal policy
- markov chain
- dynamical systems
- multiagent systems
- particle filter
- path finding
- initial state
- simulated annealing
- action space
- mobile robot
- dynamic programming
- learning algorithm