Solutions to finite horizon cost problems using actor-critic reinforcement learning.
Ivo GrondmanHao XuSarangapani JagannathanRobert BabuskaPublished in: IJCNN (2013)
Keyphrases
- reinforcement learning
- finite horizon
- actor critic
- optimal policy
- markov decision processes
- markov decision process
- function approximation
- approximate dynamic programming
- average cost
- multi agent
- temporal difference
- infinite horizon
- optimal control
- state space
- approximate solutions
- average reward
- policy gradient
- policy iteration
- learning algorithm
- fixed point
- machine learning
- decision problems
- dynamic programming
- action space
- function approximators
- reinforcement learning methods
- rl algorithms
- optimal solution