Approximate Dynamic Programming with (min; +) linear function approximation for Markov decision processes.
Chandrashekar LakshminarayananShalabh BhatnagarPublished in: CDC (2014)
Keyphrases
- approximate dynamic programming
- function approximation
- reinforcement learning
- markov decision processes
- policy iteration
- model free
- function approximators
- temporal difference
- average cost
- factored mdps
- optimal policy
- dynamic programming
- state space
- reinforcement learning algorithms
- finite state
- actor critic
- policy evaluation
- decision processes
- multi agent
- infinite horizon
- partially observable
- average reward
- machine learning
- markov decision process
- supervised learning
- control policy
- search algorithm
- continuous state
- learning algorithm
- linear program
- dynamical systems
- optimal solution
- markov decision problems
- policy gradient