Successive Convex Approximation Based Off-Policy Optimization for Constrained Reinforcement Learning.
Chang TianAn LiuGuang HuangWu LuoPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- saddle point
- concave convex procedure
- convex functions
- efficient computation
- risk minimization
- multi agent
- optimization problems
- optimization algorithm
- function approximation
- convex optimization
- convex programming
- convex relaxation
- quasiconvex
- closed form
- markov decision processes
- machine learning
- global optimization
- convex hull
- convex optimization problems
- reinforcement learning algorithms
- quadratic program
- state space
- efficient optimization
- alternating optimization
- minimize a cost function
- action selection
- piecewise linear
- constrained optimization
- error bounds
- optimization method
- objective function