Reinforcement Learning Algorithms Based on mGA and EA with Policy Iterations.
Changming YinLiyun LiHanxing WangPublished in: LSMS (1) (2007)
Keyphrases
- reinforcement learning algorithms
- reinforcement learning problems
- policy search
- partially observable environments
- reinforcement learning
- reward function
- markov decision processes
- evolutionary algorithm
- state space
- policy gradient
- model free
- optimal policy
- eligibility traces
- total reward
- markov games
- temporal difference
- reinforcement learning methods
- markov decision process
- function approximators
- learning algorithm
- action selection
- stochastic games
- markov decision problems
- function approximation
- policy iteration
- dynamic environments
- infinite horizon
- supervised learning
- optimal solution
- inverse reinforcement learning
- dynamic programming
- average reward
- partially observable
- policy evaluation
- multiagent reinforcement learning
- long run
- reward shaping