Stochastic Bilevel Optimization with Lower-Level Contextual Markov Decision Processes.
Vinzenz ThomaBarna PasztorAndreas KrauseGiorgia RamponiYifan HuPublished in: CoRR (2024)
Keyphrases
- lower level
- markov decision processes
- higher level
- low level
- upper level
- optimality conditions
- finite state
- high level
- optimal policy
- state space
- policy iteration
- transition matrices
- reinforcement learning
- dynamic programming
- decision theoretic planning
- factored mdps
- bilevel programming
- infinite horizon
- partially observable
- reachability analysis
- planning under uncertainty
- reinforcement learning algorithms
- state and action spaces
- optimization algorithm
- average reward
- action space
- average cost
- optimization problems
- action sets
- markov decision process
- model based reinforcement learning
- decision diagrams
- continuous state spaces
- finite horizon
- monte carlo
- stochastic shortest path