A Linear Approximation of the Value Function of an Approximate Dynamic Programming Approach for the Ship Scheduling Problem.
Kazuhiro KobayashiPublished in: LION (2010)
Keyphrases
- linear approximation
- approximate dynamic programming
- policy iteration
- scheduling problem
- markov decision processes
- reinforcement learning
- fixed point
- linear program
- model free
- optimal policy
- least squares
- dynamic programming
- basis functions
- control policy
- markov decision process
- average cost
- finite state
- average reward
- temporal difference
- convergence rate
- step size
- special case
- lower bound
- infinite horizon
- optimal control
- linear programming
- state space
- feature vectors