Controlled exploration of state space in off-line ADP and its application to stochastic shortest path problems.
Nikolaos E. PratikakisMatthew J. RealffJay H. LeePublished in: Comput. Chem. Eng. (2009)
Keyphrases
- shortest path problem
- state space
- shortest path
- interval data
- single source
- heuristic search
- state transition
- combinatorial optimization problems
- reinforcement learning
- continuous state spaces
- stochastic domains
- dynamic programming
- markov decision processes
- directed graph
- multiple objectives
- continuous time markov process
- optimal policy
- particle filter
- reinforcement learning algorithms
- stochastic model
- dynamical systems
- state variables
- markov chain
- planning problems
- learning automata
- stochastic processes
- initial state
- belief state
- action selection
- monte carlo
- action space
- directed acyclic graph
- genetic algorithm