Provably Efficient Convergence of Primal-Dual Actor-Critic with Nonlinear Function Approximation.
Jing DongLi ShenYinggan XuBaoxiang WangPublished in: AAMAS (2023)
Keyphrases
- function approximation
- primal dual
- actor critic
- reinforcement learning
- convergence rate
- temporal difference
- gradient method
- linear programming
- policy gradient
- linear program
- temporal difference learning
- approximate dynamic programming
- radial basis function
- convex optimization
- function approximators
- learning tasks
- model free
- step size
- genetic algorithm
- optimal control
- evolutionary algorithm
- optimal solution
- dynamic programming