|
| |
|
|
Stochastic Dynamic Programming Problems: Theory and Applications.
Lendel, Gabriel ; Sladký, Karel (advisor) ; Lachout, Petr (referee)
Title: Stochastic Dynamic Programming Problems: Theory and Applications Author: Gabriel Lendel Department: Department of Probability and Mathematical Statistics Supervisor: Ing. Karel Sladký CSc. Supervisor's e-mail address: sladky@utia.cas.cz Abstract: In the present work we study Markov decision processes which provide a mathematical framework for modeling decision-making in situations where outcomes are partly random and partly under the control of a decision maker. We study iterative procedures for finding policy that is optimal or nearly optimal with respect to the selec- ted criteria. Specifically, we mainly examine the task of finding a policy that is optimal with respect to the total expected discounted reward or the average expected reward for discrete or continuous systems. In the work we study policy iteration algorithms and aproximative value iteration algorithms. We give numerical analysis of specific problems. Keywords: Stochastic dynamic programming, Markov decision process, policy ite- ration, value iteration
|
|
|
Stochastic Dynamic Programming Problems: Theory and Applications.
Lendel, Gabriel ; Sladký, Karel (advisor) ; Lachout, Petr (referee)
Title: Stochastic Dynamic Programming Problems: Theory and Applications Author: Gabriel Lendel Department: Department of Probability and Mathematical Statistics Supervisor: Ing. Karel Sladký CSc. Supervisor's e-mail address: sladky@utia.cas.cz Abstract: In the present work we study Markov decision processes which provide a mathematical framework for modeling decision-making in situations where outcomes are partly random and partly under the control of a decision maker. We study iterative procedures for finding policy that is optimal or nearly optimal with respect to the selec- ted criteria. Specifically, we mainly examine the task of finding a policy that is optimal with respect to the total expected discounted reward or the average expected reward for discrete or continuous systems. In the work we study policy iteration algorithms and aproximative value iteration algorithms. We give numerical analysis of specific problems. Keywords: Stochastic dynamic programming, Markov decision process, policy ite- ration, value iteration
|
|
| |
guest ::
