Instructors: Dr. Alex Leong
Event type:
Lecture (with exercise)
Org-unit: Elektrotechnik und Informationstechnik
Displayed in timetable as:
K.048.92042
Language of instruction:
Englisch
Min. | Max. participants:
- | -
Requirements and recommendations:
\section{Kurzbeschreibung}
Dynamic programming is a method for solving decision making problems
consisting of a number of stages, by breaking down the problem into simpler
sub-problems. These methods have wide applicability in areas such as
optimization, control, communications, and machine learning. This course will
cover the modelling and solution of problems of sequential decision making
under uncertainty.
We will consider problems with both a finite and an infinite number of stages,
as well as cases with perfect and imperfect observations of the system.
Numerical techniques for solving these problems will be described, including
suboptimal methods for when the state and/or action spaces are large.
\section{Inhalt}
% Inhaltsangabe.
Topics to be covered in this course will include:
\begin{itemize}
\item The dynamic programming principle and dynamic programming algorithm
\item Problems with perfect state information
\item Problems with imperfect state information
\item Infinite horizon problems
\item Suboptimal methods and approximate dynamic programming
\end{itemize}
Applications to problems in control, communications, signal processing and
machine learning, including current research, will be given throughout the
course.
%----------------------------------------------
\section{Lernergebnis}
%beschreibt Lernegebnisse und KOMPETENZEN, die Studierende in dieser
%Veranstaltung erwerben können.
% Sollte hier stehen
After attending this course, students will have understood the basics of
dynamic programming and stochastic control. Students will learn the dynamic
programming optimality principle and how it can be used to solve multi-stage
decision making problems. They will learn how to formulate and solve, using
dynamic programming, problems in different areas such as control,
communications, signal processing, and machine learning.
---------------------------------------------
\section{Methodik}
mit welcher Methodik? Sollte hier stehen!
Lectures and exercises
---------------------------------------------
\section{Vorkenntnisse}
welche Vorkenntnise werden vorausgesetzt? Sollte hier stehen.
Basic knowledge on control of discrete-time systems, e.g. as covered in the
course Regelungstechnik A - Automatic Control. An introductory course on
probability and random processes, e.g. the course Stochastik f\"{u}r
Ingenieure.
----------------------------------------------
\section{Prüfung}
% Welche Prüfungsmodalitäten sind vorgesehen?
% M und S werden als Abkürzung für mündliche btw. schriftliche Prüfung
% verstanden! Sonst freier Text, mit // für D/E Unterscheidung
Written exam of 2 hours duration.
%----------------------------------------------
3
% \section{Sprache}
% Typischerweise in Excel.
% Sprache der Lehrveranstaltung
% Abkürzungen: D, E, und D+E . Sonst freien Text
% D
%----------------------------------------------
\section{Materialien}
% welche Materialien kommen zum Einsatz?
The main text will be:
\\D. Bertsekas, Dynamic Programming and Optimal Control, Vol I, 3rd Ed, Athena
Scientific, 2005
\\Some other material will be taken from:
\\D. Bertsekas, Dynamic Programming and Optimal Control, Vol II, 4th Ed,
Athena Scientific, 2012
\\M. Puterman, Markov Decision Processes, John Wiley and Sons, 1994
\\B. Anderson and J. Moore, Optimal Filtering, Prentice-Hall, 1979,
\\and various research papers.
%----------------------------------------------
\section{Bemerkung}
% irgendwelche Bemerkungen?
%%============================================
|