Choosing the cost vector of the linear programming approach to approximate dynamic programming.
Daniela Pucci de FariasThéophane WeberPublished in: CDC (2008)
Keyphrases
- approximate dynamic programming
- linear programming
- linear program
- average cost
- dynamic programming
- stochastic dynamic programming
- np hard
- reinforcement learning
- factored mdps
- optimal solution
- objective function
- control policy
- step size
- genetic algorithm
- markov decision processes
- total cost
- policy iteration
- long run
- data mining
- infinite horizon
- finite number
- function approximation
- optimal policy
- graphical models
- special case