Evo-RL: evolutionary-driven reinforcement learning.
Ahmed HallawaThorsten BornAnke SchmeinkGuido DartmannArne PeineLukas MartinGiovanni IaccaA. E. EibenGerd AscheidPublished in: GECCO Companion (2021)
Keyphrases
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- model free
- state space
- control problems
- genetic algorithm
- multi agent
- data driven
- optimal policy
- evolutionary computation
- markov decision processes
- learning process
- temporal difference
- direct policy search
- learning problems
- approximate dynamic programming
- continuous state
- rl algorithms
- temporal difference learning
- learning algorithm
- machine learning
- transfer learning
- neural network
- continuous state and action spaces
- action space
- learned knowledge
- evolutionary algorithm
- action selection
- markov decision problems
- adaptive control
- autonomous learning
- dynamic programming
- state and action spaces
- partially observable domains
- supervised learning
- dynamical systems