Stochastic Reinforcement Learning for Continuous Actions in Dynamic Environments.
Syed Naveed Hussain ShahDean Frederick HougenPublished in: FLAIRS Conference (2020)
Keyphrases
- dynamic environments
- action space
- reinforcement learning
- single agent
- continuous state spaces
- state and action spaces
- direct policy search
- state space
- stochastic processes
- action selection
- continuous action
- perceptual aliasing
- learning agent
- policy search
- markov decision processes
- path planning
- autonomous agents
- reinforcement learning algorithms
- continuous state
- mobile robot
- stochastic approximation
- state action
- plan execution
- reinforcement learning agents
- agent systems
- partially observable
- function approximation
- reward function
- sensory information
- stochastic domains
- changing environment
- collision avoidance
- real environment
- multi agent
- autonomous systems
- function approximators
- reinforcement learning methods
- optimal control
- dynamic programming
- markov decision process
- temporal difference
- adaptive control
- potential field
- agent based systems
- state transition
- simultaneous localization and mapping