Reinforcement Learning-based control using Q-learning and gravitational search algorithm with experimental validation on a nonlinear servo system.
Iuliu Alexandru ZamfiracheRadu-Emil PrecupRaul-Cristian RomanEmil M. PetriuPublished in: Inf. Sci. (2022)
Keyphrases
- reinforcement learning
- gravitational search algorithm
- function approximation
- stability analysis
- reinforcement learning algorithms
- state space
- action selection
- model free
- optimal control
- adaptive control
- control policy
- optimal policy
- learning algorithm
- convergence rate
- swarm intelligence
- dynamic programming
- machine learning
- markov decision processes
- multi objective
- control system
- control strategy
- temporal difference
- parameter identification
- search space
- control parameters
- nonlinear systems