Enabling surrogate-assisted evolutionary reinforcement learning via policy embedding.
Lan TangXiaxi LiJinyuan ZhangGuiying LiPeng YangKe TangPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- action selection
- markov decision process
- partially observable environments
- policy gradient
- reinforcement learning problems
- state space
- markov decision processes
- markov decision problems
- reinforcement learning algorithms
- actor critic
- function approximation
- reward function
- policy iteration
- action space
- function approximators
- control policies
- partially observable
- state and action spaces
- control policy
- policy evaluation
- partially observable markov decision processes
- model free
- dynamic programming
- long run
- genetic algorithm
- approximate dynamic programming
- evolutionary optimization
- state action
- vector space
- learning algorithm
- temporal difference
- evolutionary computation
- average cost
- infinite horizon
- average reward
- decision problems
- reinforcement learning methods
- rl algorithms
- continuous state spaces
- continuous state
- inverse reinforcement learning
- partially observable domains
- state dependent
- learning agent
- control problems
- optimal control
- learning process
- multi agent
- machine learning
- graph embedding
- nonlinear dimensionality reduction
- genetic programming
- low dimensional
- neural network