Feature Selection Using Regularization in Approximate Linear Programs for Markov Decision Processes
Marek PetrikGavin TaylorRonald ParrShlomo ZilbersteinPublished in: CoRR (2010)
Keyphrases
- markov decision processes
- linear program
- factored mdps
- dynamic programming
- linear programming
- average cost
- policy evaluation
- state space
- stationary policies
- optimal policy
- policy iteration
- finite state
- planning under uncertainty
- reinforcement learning
- stochastic programming
- simplex method
- transition matrices
- column generation
- partially observable
- reinforcement learning algorithms
- decision processes
- markov games
- action space
- decision theoretic planning
- markov decision process
- stochastic games
- optimal solution
- primal dual
- action sets
- multistage
- support vector
- multi objective
- markov decision problems
- np hard
- context specific
- average reward
- partially observable markov decision processes
- machine learning