Constrained episodic reinforcement learning in concave-convex and knapsack settings.
Kianté BrantleyMiroslav DudíkThodoris LykourisSobhan MiryoosefiMax SimchowitzAleksandrs SlivkinsWen SunPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- piecewise linear
- dynamic programming
- convexity properties
- convex functions
- knapsack problem
- convex concave
- function approximation
- markov decision processes
- saddle point
- reinforcement learning algorithms
- state space
- optimal policy
- objective function
- convex hull
- convex optimization
- optimal control
- upper bound
- temporal difference
- feasible solution
- transition model
- learning algorithm
- multi agent reinforcement learning
- lower bound
- convex relaxation
- optimal solution
- robotic control
- action selection
- learning classifier systems
- globally optimal
- evolutionary algorithm
- machine learning
- transfer learning
- supervised learning
- learning process