Approximating a deep reinforcement learning docking agent using linear model trees.
Vilde B. GjærumElla-Lovise H. RørvikAnastasios M. LekkasPublished in: ECC (2021)
Keyphrases
- linear model
- reinforcement learning
- regression trees
- multi agent
- action selection
- least squares
- linear models
- regression model
- learning agent
- multi agent systems
- autonomous agents
- decision trees
- additive model
- intelligent agents
- autonomous learning
- agent receives
- nonlinear models
- function approximation
- multiagent systems
- linear transformation
- reward function
- learning agents
- multiple agents
- learning capabilities
- learning algorithm
- multi agent environments
- dynamic environments
- mobile agents
- single agent
- partially observable
- state space
- optimal control
- machine learning
- software agents
- markov decision process
- markov decision processes
- state action
- linear transformations
- state abstraction
- semi parametric
- exploration strategy
- computer vision