On linear and super-linear convergence of Natural Policy Gradient algorithm.
Sajad KhodadadianPrakirt Raj JhunjhunwalaSushil Mahavir VarmaSiva Theja MaguluriPublished in: Syst. Control. Lett. (2022)
Keyphrases
- learning algorithm
- convergence rate
- dynamic programming
- worst case
- cost function
- search space
- np hard
- policy gradient
- monte carlo
- simulated annealing
- least squares
- state space
- computational complexity
- dynamic environments
- support vector machine svm
- optimal solution
- decision problems
- objective function
- recursive least squares
- neural network