skrl: Modular and Flexible Library for Reinforcement Learning.
Antonio Serrano-MuñozNestor Arana-ArexolaleibaDimitrios ChrysostomouSimon BøghPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- highly flexible
- function approximation
- state space
- multi agent
- learning process
- model free
- neural network
- learning algorithm
- modular neural networks
- modular structure
- markov decision processes
- robotic control
- function approximators
- control problems
- learning capabilities
- temporal difference
- lightweight
- dynamic programming
- multiscale
- genetic algorithm