A Best-of-Both-Worlds Algorithm for Constrained MDPs with Long-Term Constraints.
Jacopo GermanoFrancesco Emanuele StradiGianmarco GenaltiMatteo CastiglioniAlberto MarchesiNicola GattiPublished in: CoRR (2023)
Keyphrases
- similarity measure
- long term
- markov decision processes
- linear programming
- np hard
- constrained optimization
- short term
- optimization algorithm
- detection algorithm
- model based reinforcement learning
- multi agent
- multiple constraints
- constrained problems
- convex hull
- bayesian networks
- monte carlo
- expectation maximization
- decision trees
- worst case
- probabilistic model
- dynamic programming
- reinforcement learning
- preprocessing
- optimal solution