Improved Stochastic Synapse Reinforcement Learning for Continuous Actions in Sharply Changing Environments.
Syed Naveed Hussain ShahDean Frederick HougenPublished in: IJCNN (2020)
Keyphrases
- changing environment
- reinforcement learning
- action space
- continuous state spaces
- dynamic environments
- state space
- stochastic processes
- action selection
- direct policy search
- control policies
- sensory information
- autonomous agents
- partially observable
- perceptual aliasing
- stochastic approximation
- markov decision processes
- state and action spaces
- multiagent reinforcement learning
- continuous domains
- continuous state and action spaces
- multi agent systems
- continuous state
- partially observable domains
- multiagent systems
- function approximation
- reinforcement learning algorithms
- state action
- learning agent
- machine learning
- model free
- continuous action
- behavioural cloning
- optimal control problems
- real time
- stochastic domains
- partial observability
- learning algorithm
- multi agent
- cooperative
- state transition