A Forward-backward Algorithm for Stochastic Control Problems - Using the Stochastic Maximum Principle as an Alternative to Dynamic Programming.
Stephan E. LudwigJustin A. SirignanoRuojun HuangGeorge PapanicolaouPublished in: ICORES (2012)
Keyphrases
- dynamic programming
- forward backward
- control problems
- reinforcement learning
- continuous state spaces
- monte carlo
- learning algorithm
- np hard
- state space
- linear programming
- optimal control
- hidden markov models
- k means
- objective function
- particle swarm optimization
- optimization algorithm
- mathematical model
- convergence rate
- optimal solution
- stochastic processes
- locally optimal