Collaborative Exploration and Reinforcement Learning between Heterogeneously Skilled Agents in Environments with Sparse Rewards.
Alain AndresEsther Villar-RodriguezAritz D. MartinezJavier Del SerPublished in: IJCNN (2021)
Keyphrases
- reinforcement learning
- action selection
- multi agent environments
- multi agent
- exploration strategy
- learning agents
- single agent
- multi agent systems
- dynamic environments
- markov decision processes
- reinforcement learning agents
- multi agent reinforcement learning
- cooperative
- learning agent
- autonomous agents
- function approximation
- highly dynamic
- reinforcement learning algorithms
- multiagent systems
- state space
- complex environments
- robocup soccer
- learning capabilities
- intelligent agents
- open systems
- model based reinforcement learning
- high dimensional
- agent receives
- active exploration
- learning algorithm
- collaborative learning
- model free
- multiagent learning
- software agents
- temporal difference
- reward function
- multiple agents
- dynamic programming
- optimal policy
- agent model
- function approximators
- decision making
- complex domains
- learning process
- autonomous entities
- agent learns
- machine learning
- sparse representation
- resource allocation
- autonomous learning
- partial observability
- stochastic games
- reward signal
- action space
- partially observable markov decision processes