Training a RoboCup Striker Agent via Transferred Reinforcement Learning.
Warren Blair Watkinson IITracy CampPublished in: RoboCup (2018)
Keyphrases
- multi agent
- reinforcement learning
- robocup soccer
- action selection
- multiagent systems
- robot soccer
- function approximation
- learning agent
- multi agent systems
- single agent
- robotic soccer
- state space
- learning algorithm
- multiple agents
- partially observable
- autonomous agents
- intelligent agents
- state abstraction
- optimal policy
- supervised learning
- reinforcement learning agents
- training process
- machine learning
- reinforcement learning algorithms
- markov decision process
- real robot
- exploration strategy
- learning capabilities
- agent receives
- reward function
- test bed
- markov decision processes
- training examples
- training samples
- training set
- decision making
- agent model
- agent systems
- transfer learning
- mobile agents
- dynamic environments
- learning process
- artificial neural networks
- multiagent reinforcement learning
- cooperative