Experiments on the use of option policies in reinforcement learning.
Letícia Maria FriskeCarlos H. C. RibeiroPublished in: J. Intell. Fuzzy Syst. (2002)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- control policies
- fitted q iteration
- reinforcement learning agents
- markov decision processes
- reward function
- state space
- hierarchical reinforcement learning
- function approximation
- partially observable markov decision processes
- cooperative multi agent systems
- decision problems
- policy gradient methods
- learning algorithm
- reinforcement learning algorithms
- total reward
- dynamic programming
- control policy
- model free
- markov decision problems
- temporal difference
- finite state
- robotic control
- revenue management
- state abstraction
- supervised learning
- macro actions
- infinite horizon
- continuous state
- average reward
- sufficient conditions
- learning agent
- approximate policy iteration
- natural actor critic
- long run