Error bounds for constant step-size Q-learning.
Carolyn L. BeckR. SrikantPublished in: Syst. Control. Lett. (2012)
Keyphrases
- error bounds
- step size
- convergence rate
- cost function
- reinforcement learning
- learning rate
- convergence speed
- theoretical analysis
- worst case
- faster convergence
- learning algorithm
- state space
- function approximation
- hessian matrix
- variable step size
- adaptive filter
- temporal difference
- model free
- stochastic gradient descent
- line search
- multi agent
- wavelet coefficients
- optimal policy
- steepest descent method
- action selection
- gradient method
- polynomial time approximation
- wavelet synopses
- face recognition