Successive Convex Approximation Based Off-Policy Optimization for Constrained Reinforcement Learning.
Chang TianAn LiuGuan HuangWu LuoPublished in: IEEE Trans. Signal Process. (2022)
Keyphrases
- reinforcement learning
- saddle point
- concave convex procedure
- convex programming
- convex functions
- optimization algorithm
- function approximation
- optimization process
- model free
- quasiconvex
- convex relaxation
- constrained optimization
- optimization problems
- global optimization
- error bounds
- risk minimization
- closed form
- convex optimization
- reinforcement learning algorithms
- convex hull
- approximation error
- quadratic program
- neural network
- stationary points
- piecewise linear
- efficient optimization
- optimal policy
- semi definite programming
- min sum
- machine learning