Symbolic Dynamic Programming for Continuous State MDPs with Linear Program Transitions.
Jihwan JeongParth JaggiScott SannerPublished in: IJCAI (2021)
Keyphrases
- linear program
- continuous state
- dynamic programming
- dec pomdps
- markov decision processes
- reinforcement learning
- action space
- partially observable markov decision processes
- finite state
- state space
- state transition
- linear programming
- optimal policy
- policy search
- continuous state and action spaces
- state dependent
- continuous state spaces
- control policies
- robot navigation
- approximate dynamic programming
- multistage
- column generation
- planning problems
- markov decision problems
- average cost
- optimal solution
- optimal control
- function approximation
- state transitions
- objective function
- policy iteration
- machine learning
- partially observable
- reward function
- np hard
- dynamical systems
- markov chain
- learning algorithm
- convex optimization
- markov decision process
- finite number