Contractive approximations in average Markov decision chains driven by a risk-seeking controller.
Gustavo Portillo-RamírezRolando Cavazos-CadenaHugo Cruz-SuárezPublished in: Math. Methods Oper. Res. (2023)
Keyphrases
- markov decision chains
- risk sensitive
- average cost
- optimal control
- control system
- utility function
- closed loop
- real time
- long run
- markov decision processes
- dynamic programming
- control policy
- approximation methods
- initial state
- model free
- infinite horizon
- machine learning
- control law
- fixed point
- finite state
- pid controller
- finite number
- total cost
- expected utility
- control method
- control algorithm
- risk management
- decision problems
- fractal image coding
- reinforcement learning