Parallel Implementation of Reinforcement Learning Q-Learning Technique for FPGA.
Lucileide M. D. Da SilvaMatheus F. TorquatoMarcelo A. C. FernandesPublished in: IEEE Access (2019)
Keyphrases
- parallel implementation
- reinforcement learning
- function approximation
- parallel architecture
- reinforcement learning algorithms
- model free
- state space
- hardware implementation
- learning algorithm
- parallel implementations
- parallel computation
- field programmable gate array
- temporal difference learning
- action selection
- multi agent
- markov decision processes
- multi agent reinforcement learning
- parallel computers
- optimal policy
- continuous state
- reinforcement learning methods
- general purpose
- dynamic programming
- machine learning
- low cost
- mobile robot
- stochastic approximation
- temporal difference
- single chip
- real robot
- markov decision problems
- actor critic
- policy search
- td learning
- eligibility traces
- continuous state and action spaces