Approximate dynamic programming using support vector regression.
Brett BethkeJonathan P. HowAsuman E. OzdaglarPublished in: CDC (2008)
Keyphrases
- approximate dynamic programming
- linear program
- reinforcement learning
- stochastic dynamic programming
- dynamic programming
- step size
- policy iteration
- control policy
- average cost
- factored mdps
- linear programming
- markov decision processes
- state space
- least squares
- optimal policy
- finite number
- feature extraction
- decision making