DEMO: Self-Preserving Genetic Algorithms vs. Safe Reinforcement Learning in Discrete Action Spaces.
Preston K. RobinetteNathaniel P. HamiltonTaylor T. JohnsonPublished in: ICCPS (2023)
Keyphrases
- action space
- reinforcement learning
- continuous state
- continuous action
- continuous state spaces
- genetic algorithm
- state space
- policy search
- markov decision processes
- state and action spaces
- real valued
- stochastic processes
- action selection
- control policies
- reinforcement learning methods
- evolutionary algorithm
- finite state
- state action
- robot navigation
- model free
- temporal difference
- skill learning
- function approximation
- simulated annealing
- learning algorithm
- learning classifier systems
- dynamic programming
- optimal policy
- state dependent
- planning problems
- partially observable markov decision processes
- single agent
- reinforcement learning algorithms
- supervised learning