Book of Modules for the Elite Master’s Programme

Scientific Computing
Faculty for Mathematics, Physics and Computer Science, University of Bayreuth, Germany

October 30, 2019

Preface
This Book of Modules describes all currently admissible modules for the elite master’s programme Scientific Computing and
their assignment to the module areas decribed in the examination regulations ”Studien- und Prüfungsordnung”. Each module
description contains the frequency (and sometimes the term) with which the corresponding course is offered; the list of courses
actually provided during the current semester is published via Campus Online. The coordinator of this master’s programme will
be glad to help you selecting suitable courses to satisfy the requirements formulated in the examination regulations.
The executive committee of the elite master’s programme October 30, 2019

Contents
1 Module Section A: Numerical Mathematics 3
1.1 Module A1: Numerical Methods for Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Elective Modules A2: Advanced Topics in Numerical Mathematics . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Module Section B: Modeling and Simulation 9

2.1 Module B1: Applied Functional Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Elective Modules B2: Modeling and Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Module B3: Industrial Internship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4 Module B4: Modeling and Status Seminar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

3 Module Section C: High-Performance Computing 24

3.1 Elective Modules C1: High-Performance Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Module C2: Practical Course on Parallel Numerical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4 Module Section D: Scientific Computing 32

4.1 Elective Modules D1: Complexity Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2 Module D2: Special Skills in Scientific Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Module Section E: Soft Skills 44

6 Module Section F: Master’s Thesis 45

7 Recommended Curriculum 46

2
1 Module Section A: Numerical Mathematics

1.1 Module A1: Numerical Methods for Differential Equations

Title Numerical Methods for Differential Equations

Module Label A1

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chairs of Applied and Numerical Analysis, Applied Mathematics, Scientific Computing

Learning Outcomes This module lays ground to many modules of this programme. It should be attended during the
first semester in order to be able to

• understand the way numerical algorithms for the solution of ordinary and partial
differential equations work

• choose a suitable algorithm for a given class of ordinary and partial differential equations

• adapt standard algorithms to new problems

• implement the algorithms discussed in the lecture in MATLAB or in a higher programming
language.

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content • One step methods for ordinary differential equations (convergence, Runge-Kutta- and
extrapolation methods, methods for stiff differential equations, step size control)

• multi step schemes for ODEs

• finite difference methods for partial differential equations (transport, Poisson, heat, wave
equation)

• conforming finite element methods for elliptic PDEs (Galerkin method, convergence)

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 6

Work Load 4 h lectures plus 2 h post-processing per week = 90 h;

2 h discussion sections plus 3 h preparation/post-processing = 75 h;
15 h preparation for exam, in total: 180 h

Recommended Introduction to Numerical Mathematics, Higher Skills in Numerical Mathematics; helpful but not
Prerequisites required: Introduction to Ordinary Differential Equations, Introduction to Advanced Analysis

Grading Oral or written exam; active participation in the tutorials

Frequency Every year during winter term

3
1.2 Elective Modules A2: Advanced Topics in Numerical Mathematics

Title Numerical Methods for General Types of PDEs

Module Label A2.1

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chairs of Applied and Numerical Analysis, Scientific Computing

Learning Outcomes • Understanding the way numerical algorithms for the solution of special partial differential
equations work

• Ability to choose a suitable discretization technique for a given partial differential equation
• Ability to choose a suitable algorithm
• Ability to implement the algorithms discussed in the lecture in a higher programming
language

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content This module is the continuation of the module A1: Numerical Methods for Differential Equations.
It is focused on the numerical solution of more general types of partial differential equations
arising from realistic applications such as fluid dynamics, electromagnetism, structural
mechanics, etc. These require special discretization techniques:

• non-conforming and mixed finite element methods

• finite element methods for (Navier-)Stokes equations

• finite volume methods

• discontinuous Galerkin methods
• edge elements
Special topics:

• adaptivity
• smoothed particle hydrodynamics
• mortar methods

• level-set methods

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended A1: Numerical Methods for Differential Equations

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

4
 Title Constructive Approximation Methods
Module Label A2.2

Module Type Lecture

Area of Research Advanced Analysis and Applications / Numerical Mathematics

Responsible Chair of Applied and Numerical Analysis

Learning Outcomes By the end of the course, a successful student should be able to

• explain the most important concepts of modern, multivariate approximation methods

• explain the problems inherent to the reconstruction of multivariate functions from
scattered data

• prove and analyse the existence, the uniqueness, the computability and the quality of
discrete reconstruction techniques

• explain and implement the associated numerical schemes

• understand the underlying mathematical theory

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content • Jackson- and Bernstein theorems for classical univariate polynomial approximation
• Multivariate reconstruction methods based upon radial basis functions, moving
least-squares and partition of unity methods

• Error and stability analysis of multivariate reconstruction methods

• Development and implementation of efficient algorithms for such reconstruction methods
• Optimal recovery for generalised interpolation with application to solving partial differential
equations

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended Introduction to Advanced Analysis, Introduction to Numerical Mathematics

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

5
 Title Mathematical Control Theory
Module Label A2.3

Module Type Lecture

Area of Research Optimization / Advanced Analysis and Applications

Responsible Chair of Applied Mathematics

Learning Outcomes
• knowledge of methods and concepts of mathematical control theory
• ability to solve selected problems from mathematical control theory

• ability to apply these solution strategies to practical problem formulations

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content
• definition and classification of control systems
• qualitative analysis of control systems

• methods for controller design, e.g.

– methods from linear algebra
– methods from optimal control
– methods based on Lyapunov functions

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended Introduction to Numerical Mathematics, A1: Numerical Methods for Differential Equations
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

6
 Title Nonlinear Optimization
Module Label A2.4

Module Type Lecture

Area of Research Optimization

Responsible Chairs of Applied Mathematics, Scientific Computing

Learning Outcomes
• understanding of optimatlity conditions for nonlinear optimization
• understanding of the most important algorithms for the numerical solution of nonlinear
optimization problems

• ability to model and solve given practical problems in nonlinear optimization

• ability to use and develop software for nonlinear optimzation

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content
• modeling of nonlinear optimization problems

• advanced algorithms for unconstrained optimization

• optimality conditions for nonlinear optimization problems
• algorithms for constrained optimization

• outlook on further problem classes

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended Introduction to Numerical Mathematics, Higher Skills in Numerical Mathematics

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

7
 Title Optimization of Differential Equations
Module Label A2.5

Module Type Lecture

Area of Research Numerical Mathematics / Optimization

Responsible Chairs of Applied Mathematics, Scientific Computing

Learning Outcomes • ability to derive and analyse optimality conditions

• ability to apply theory to concrete applications

• ability to solve the arising problems numerically

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content • Introductory examples and concepts

• General existence theory and first order optimality conditions

• Linear-quadratic problems
• Introduction to some non-linear problems
• Basic numerical methods
• Examples from applications

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended B1: Applied Functional Analysis, A1: Numerical Methods for Differential Equations
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

8
2 Module Section B: Modeling and Simulation

2.1 Module B1: Applied Functional Analysis

Title Applied Functional Analysis

Module Label B1

Module Type Lecture

Area of Research Partial Differential Equations

Responsible Chairs of Applied and Numerical Analysis, Applied Mathematics, Scientific Computing,
Nonlinear Analysis and Mathematical Physics

Learning Outcomes This module lays ground to many modules of this programme. It should be attended during the
first semester. By the end of the course, a successful student should

• know the basic solution spaces and understand their uses in the theory of partial
differential equations;

• master the concept of a weak solution;

• be able to apply functional analysis methods to problems in partial differential equations;
• understand how functional analysis concepts develop out of applications.
In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content Basic solution spaces and methods from functional analysis which are used for analysing partial
differential equations, in particular

• Sobolev spaces, embedding theorems

• weak solutions of elliptic equations, Lax-Milgram lemma, Fredholm alternative
• regularity of weak solutions of elliptic equations
• spectral theory for compact operators

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended Introduction to Advanced Analysis

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year during winter term

9
2.2 Elective Modules B2: Modeling and Simulation

Title Partial Differential Equations and Integral Equations

Module Label B2.1

Module Type Lecture

Area of Research Partial Differential Equations

Responsible Chair of Nonlinear Analysis and Mathematical Physics

Learning Outcomes By the end of the course, a successful student should

• understand the origin of the treated equations in the modeling process;

• know fundamental results on existence and uniqueness of their solutions;

• understand qualitative properties of their solutions;

• master key methods of their analysis.
In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content Existence, uniqueness, and properties of solutions for integral equations and for various types of
partial differential equations that are eminent for modeling in the sciences, in particular

• parabolic equations

• wave equations and symmetric hyperbolic systems

• Schrödinger equations
• integral operators related to elliptic equations

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam, in total: 240 h

Recommended Introduction to Advanced Analysis, B1: Applied Functional Analysis

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

10
 Title Modeling with Differential Equations
Module Label B2.2

Module Type Lecture

Area of Research Partial Differential Equations

Responsible Chairs of Applied Mathematics, Scientific Computing

Learning Outcomes • Ability to identify suitable mathematical models for a given application
• Knowledge about important modeling principles
• Ability to apply modeling techniques to basic practical applications

In contrast to students who have passed bachelor module C1, students of this modules can
apply the previous techniques more autonomously and can relate them to formerly acquired
advanced skills.

Content • General modeling principles

• Mathematical Models based on Ordinary Differential Equations from, e.g.,
– Mathematical Biology
– Mechanics

• Mathematical Models based on Partial Differential Equations from, e.g.,

– Mathematical Physics
– Mathematical Finance

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended B1: Applied Functional Analysis, A1: Numerical Methods for Differential Equations
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

11
 Title Mathematical Modeling for Climate and Environment
Module Label B2.3

Module Type Lecture

Area of Research Partial Differential Equations

Responsible Chair of Scientific Computing

Learning Outcomes • Knowledge of important physical principles and their representation in mathematical
models for main types of climate and environmental models

• Ability to identify the key interactions between different compartments of a climate model
and to express them in mathematical form

• Ability to formulate simple environmental and climate models and skills to implement them
using e.g. Matlab

Content • Physical principles, mathematical models, and selected numerical methods in climate and
environmental sciences

• Earth system: Main components, driving forces, scales, feedbacks

• Hierarchy of climate models, regional and global focus
• Environmental modeling: Main applications and problem settings

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 2 h post-processing per week = 90 h;

120 h practical course; 30 h preparation for exam; in total: 240 h

Recommended B1: Applied Functional Analysis, A1: Numerical Methods for Differential Equations,
Prerequisites basic programming skills in Matlab

Grading Oral exam; active participation in the tutorials

Frequency Once per year, summer term

12
 Title Pattern Recognition
Module Label B2.4

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Applied Computer Science III

Learning Outcomes This course imparts advanced, systematic comprehension and methods to recognize or classify
patterns in a set of data. E. g. applications are in the fields of object recognition, recognition of
hand writing, speech, or gestures, and facial recognition.

Content Bayesian classification, Parameter estimation, Nonparametric techniques, Linear classification,

Feedforward neural networks, Feedback neural networks, Nonmetric methods, Supervised
Learn- ing, Unsupervised Learning

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1 h post-processing per week = 45 h;

60 h practical course; 15 h preparation for exam; in total: 120 h

Recommended none
Prerequisites

Grading Oral exam; For the admission to the written exam a vivid participation in the exercises is
required.

Frequency Once per year, winter term

13
 Title Computational Biophysics
Module Label B2.5

Module Type Lecture with practical course

Area of Research Physics / Chemistry

Responsible Biofluid Simulation and Modeling, Theoretische Physik VI

Learning Outcomes • Understanding of the theoretical background of Density Functional Theory and Classical
Molecular Dynamics

• Practical application to example problems, e.g. for protein folding

Content This lecture gives an introduction to a set of computational tools used in modern chemistry
covering Density Functional Theory as well as classical Molecular Dynamics simulations. The
goal is an understanding of the theoretical backgrounds, the computational methods as well as
their practical application.

Duration 1 semester

Language English

Teaching Method lectures (2 h/week) and practical course

Credit Points 4

Work Load 2 h lectures plus 1 h post-processing per week = 45 h;

60 h practical course; 15 h preparation for exam; in total: 120 h

Recommended None
Prerequisites

Grading Oral exam

Frequency Every year

14
 Title Simulation of Biophysical Systems
Module Label B2.6

Module Type Lecture

Area of Research Physics

Responsible Biofluid Simulation and Modeling

Learning Outcomes • Understanding of the theoretical background in biophysical simulations

• Practical application to example problems, e.g. in blood flow

Content Biological flows play a decisive role in every living organism (blood flow in animals/humans,
water distribution in plants). The physical complexity of such flows arises from an interplay
between different components (cells, plasma, proteins). Accordingly, there exists a large variety
of computational methods to tackle such problems. This lecture will give an overview of the
different theoretical backgrounds as well as the practical advantages and disadvantages of
these methods.

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (2 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

2 h discussion sections plus 1.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended None
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Approximately every two years

15
 Title Bioinformatics: Molecular Modeling
Module Label B2.7

Module Type Lecture with practical course

Area of Research Biochemistry

Responsible Bioinformatics / Structural Biology

Learning Outcomes Knowledge on modeling and analysis of biomolecular processes such as enzymatic reactions
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content The lecture treats the theoretical foundations of molecular modeling (molecular force fields,
biomolecular electrostatics, classical and statistical mechanics), their numerical designs
(molecular dynamics simulations, energy minimization and normal mode analysis, Monte Carlo
simulations), fundamentals quantum chemical methods as well as the modeling of biochemical
reactions and ligand binding. In the practical course, various techniques (including analysis of
biomolecular structures, computation of electrostatic properties of biomolecules, normal-mode
analysis) will be exemplified by selected case studies to provide students with practical
demonstrations of these methods.

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and practical course (3 weeks)

Credit Points 4

Work Load 2 h lectures plus 1 h post-processing per week = 45 h;

60 h practical course; 15 h preparation for exam; in total: 120 h

Recommended Basic knowledge of chemistry and some biochemistry

Prerequisites

Grading Oral or written exam; active participation in the practical course

Frequency Every year during winter term

16
 Title Foundations of Bioinformatics
Module Label B2.8

Module Type Lecture with practical course

Area of Research Biochemistry

Responsible Bioinformatics / Structural Biology

Learning Outcomes Students should acquire the basics of bioinformatics and get to know them in theory and
practice.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content The lecture covers the basic bioinformatic applications. Namely, the application of various
theoretical methods in the analysis of molecular biological data in the foreground (databases
and database search, sequences and sequence alignments, phylogenetic trees) as well as
fundamentals of molecular modeling, structure prediction and drug design. In the practical
course, the students have hands-on sessions for the different methods, the use of internet tools
for sequence data analysis, web-based databases and creation of sequence alignments.
Moreover some basic introduction to molecular visualization and an introduction to the UNIX
operating system are provided.

Duration 1 semester

Language German

Teaching Method Lectures (2 h/week) and practical course (3 h/week)

Credit Points 4

Work Load 2 h lectures plus 2 h post-processing per week = 60 h

3 h practical course = 45 h; 15 h preparation for exam; in total: 120 h

Recommended Basic knowledge of chemistry and some biochemistry

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year during summer term

17
 Title Higher Strengths of Materials
Module Label B2.9

Module Type Lecture

Area of Research Engineering Science

Responsible Chair of Design and CAD

Learning Outcomes The module enables the students to calculate complex technical products by use of linear
elasticity approach. It deepens the knowledge in the area of strengths of materials calculation.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content Selected subjects in higher strengths of materials area, e.g. multiaxial stress and deformation,
theories of thin shells, mechanical vibrations with relation to typical applications.

Duration 1 semester

Language German

Teaching Method Lectures (2 h/week) and tutorials (2 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

2 h discussion sections plus 1.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended Basic engineering knowledge funding on Bachelor Engineering Science studies, especially in
Prerequisites technical mechanics and strengths of materials.

Grading Written examination (60 min)

Frequency Every year

18
 Title Computer Aided Engineering
Module Label B2.10

Module Type Lecture

Area of Research Engineering Science

Responsible Chair of Design and CAD

Learning Outcomes CAE1: ability to create CAD models and generate design proposals using optimization
algorithms.
CAE2: mastery of modern methods of calculation of statics and their application to constructive
tasks; knowledge of associated software
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content CAE1: mastery of modern calculation methods and their application to constructive tasks;
knowledge of associated software. Ability to design independently using CAD.
CAE2: theory and application of finite element methods to static problems with a focus on the
constructive point of view and modeling.

Duration 2 semesters

Language English

Teaching Method CAE1: lectures (2 h/week) and CAE2: seminar (2 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

2 h seminar plus 1.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended Basic technical understanding

Prerequisites A1: Numerical Methods for Differential Equations

Grading Written examination; For the admission to the written exam a vivid participation in the exercises
is required.

Frequency Every year

19
 Title Model Building and Simulation of Mechanical Systems
Module Label B2.11

Module Type Lecture

Area of Research Engineering Science

Responsible Chair of Design and CAD

Learning Outcomes The industrial standard CAD software CATIA enables students to create virtual models of
products. The course Higher Finite Element Analysis enables students to build-up models for
the dimensioning of complex technical products by use of sophisticated Finite Element Analysis
methods. The knowledge is used in wide areas of advanced product development.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content FEA: Handling of great and complex structures, use of shell and volume elements. Solving
non-linearity, vibration and heat transfer problems.
CATIA: Part and assembly creation; generation of technical drawings; surface modeling with
Generative Shape Design.

Duration 2 semesters

Language German

Teaching Method FEA: lectures (2 h/week), seminar (1 h/week) and CATIA: seminar (2 h/week)

Credit Points 6

Work Load FEA:

45 h lecture + follow up work
45 h seminars + preparation and follow-up work

CATIA:
60 h seminars + preparation and follow-up work

Examination: 30 h examination preparation; Module in total: 180 hours

Recommended Basic engineering knowledge funding on Bachelor Engineering Science studies, especially in
Prerequisites technical mechanics, construction design and mechanical engineering

Grading Written examination (120 min)

Frequency Every year

20
 Title Foundations of Data Management
Module Label B2.12

Module Type Lecture

Area of Research Data Management, Data Science

Responsible Chair of Theoretical Computer Science

Learning Outcomes Students will learn the mathematical foundations of data management (which includes
databases and data science). They will understand the connections between logic, expressivity,
computational complexity, and efficient algorithms in this area. They will learn the formal tools to
be able to understand and interpret recent scientific developments in the area.

Content The lecture starts with formal definitions of databases and query languages. After showing that
there is a deep connection between first-order logic and SQL when it comes to querying
relational databases, it investigates the computational complexity (or efficient algorithms) for
evaluating and analyzing SQL or first-order logic queries on databases. We then investigate
conjunctive queries as a practically relevant special case, treat their evaluation and optimization
problems, and connections with graph theory. (Knowledge of SQL is helpful but is not required.)

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h exercises plus 1 h post-processing per week = 30 h;
37.5 h preperation for the exam; in total: 120 h

Recommended Theoretische Informatik 1 or mathematical skills equivalent to those obtained in a BSc degree in
Prerequisites mathematics or physics

Grading Oral or written exam; participation in the tutorials

Frequency Every year

21
2.3 Module B3: Industrial Internship

Title Industrial Internship

Module Label B3

Module Type Internship

Area of Research Scientific Computing

Responsible Chair of Scientific Computing

Learning Outcomes By the end of the internship, a successful student should have

• gained hands-on experience in a non-academic environment

• applied mathematical techniques and techniques from computer science to real
applications

• acquired new ideas for his/her own research

• written a short report

Content Industrial applications in the area of Scientific Computing

Duration 6 weeks

Language English

Teaching Method Practical course

Credit Points 8

Work Load 6 weeks of internship = 230 h

10 h preparation of report; in total: 240 h

Recommended Basic lectures of this programme (after first year)

Prerequisites

Grading Written report of at least 10 pages (to be submitted not later than 4 weeks after the internship)

Frequency Every semester

22
2.4 Module B4: Modeling and Status Seminar

Title Modeling and Status Seminar

Module Label B4

Module Type Seminar

Area of Research All areas

Responsible Chair of Scientific Computing

Learning Outcomes Successful students can

• Modeling and numerical solution:

transfer real-world problems to a mathematical model
pave their way into a scientific topic; work in small groups
select suitable efficient numerical methods and implement them on parallel computers

• Talk:
choose and master suitable presentation techniques
speak freely about a subject and illustrate important structures instructively
answer spontaneous questions from the audience in a reliable manner

• Discussion:
phrase appropriate subject-specific questions
express constructive criticism for a talk
exploit constructive criticism for their future talks

• Handout:
expose an advanced mathematical subject briefly, concisely, and memorably in writing
efficient usage of scientific publication systems (e.g., LATEX)

Content Modeling Seminar:

• Students receive real-world projects and work (in small groups) their way into them
• Each group prepares a presentation for its subject (duration: 30–60 minutes) and talks
about it in front of the plenum

• Each group prepares and distributes a report (at least 10 pages) using a scientific text
system (e.g., LATEX)

Status Seminar:

• Each student prepares a presentation on the status of his/her studies and results of
his/her research (duration: 15–30 minutes) and talks about it in front of the plenum

For both seminars there will be a discussion on the subject and on the presentation.

Duration 4 semesters

Language English

Teaching Method Modeling seminar (1 week) and status seminar (2 days) each year

Credit Points 8

Work Load 70 h practical course and 30 h seminar each year = 200 h

40 h preparation/post-processing for seminar, in total: 240 h

Recommended At least one module of A2 and D1, respectively; C2: Practical Course on Parallel Numerical
Prerequisites Methods

Grading Oral presentation and written report of at least 10 pages (to be submitted not later than 4 weeks
after the seminar)

Frequency Each year (modeling seminar during summer break, status seminar during winter break)

23
3 Module Section C: High-Performance Computing

3.1 Elective Modules C1: High-Performance Computing

Title Algorithms and Data Structures II

Module Label C1.1

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Algorithms and Data Structures

Learning Outcomes This module teaches advanced techniques for the design and analysis of algorithms and data
structures.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content Possible topics are:

• design principles

• graph algorithms
• advanced data structures
• approximation algorithms
• parameterized algorithms

• randomized algorithms

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam; in total: 240 h

Recommended Elementary programming skills, Basic skills in the design and analysis of algorithms.
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

24
 Title Algorithms and Data Structures III
Module Label C1.2

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Algorithms and Data Structures

Learning Outcomes This module teaches specialized techniques for the design and analysis of algorithms and data
structures
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content Possible topics are:

• geometric algorithms

• algorithms for data analysis

• streaming algorithms
• external-memory algorithms

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended Advanced programming skills, Advanced skills in the design and analysis of algorithms
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

25
 Title Parallel and Distributed Systems I
Module Label C1.3

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Parallel and Distributed Systems

Learning Outcomes The goal of this course is to impart to the students basic techniques in parallel and distributed
programming. By that, special methodical competences are acquired: By understanding basic
problems such as load balancing and scalability and by learning synchronization and
communication techniques, the students are enabled to design and, with the help of
communication and thread libraries, to transform parallel algorithms into efficient parallel and
distributed programs. By that, both shared and distributed address spaces are acquired.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content • Architecture and interconnection networks for parallel systems

• Performance analysis and scalability of parallel programs

• Programming and synchronization techniques for shared address space with

multi-threading

• Coordination of parallel and distributed programs

• Application of programming techniques to complex examples from different areas

• Programming techniques for distributed address spaces and message-passing

Duration 1 semester

Language German

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 120 h in total (45 h presence, 60 h preparation/post-processing, 15 h preparation for exam)

Recommended
Prerequisites

Grading Written exam; For the admission to the written exam a vivid participation in the exercises is
required.

Frequency Every year in winter term

26
 Title Parallel and Distributed Systems II
Module Label C1.4

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Parallel and Distributed Systems

Learning Outcomes The goal of this course is to give the students a deep understanding of important techniques in
parallel and distributed programming. The emphasis lies on the acquiring of methodical and
technical competences. Based on a deep understanding of standard protocols for computer
networks such as IP or TCP/UDP, the students are enabled to design and implement distributed
programs. The course covers message-passing approaches such as MPI, passiv
communication mechanisms such as sockets, and also active mechanisms such as RPC, RMI,
or CORBA. The course also imparts design and implementation competences by applying the
techniques to a variety of examples.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content The course covers the basics of parallel and distributed systems with an emphasis on distributed
systems. Based on the first part of the course, the following topics are covered:

• Message-Passing programming (MPI)

• Important communication protocols in distributed systems
• Communication coordination and synchronization mechanisms in distributed systems
(examples: Sockets, RPC, Java RMI)

• Coordination with distributed objects (example: CORBA)

• Security aspects and mechanisms in distributed systems

Duration 1 semester

Language German

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 120 h in total (45 h presence, 60 h preparation/post-processing, 15 h preparation for exam)

Recommended C1.3: Parallel and Distributed Systems I

Prerequisites

Grading Written exam; For the admission to the written exam a vivid participation in the exercises is
required.

Frequency Every year in summer term

27
 Title High-Performance Computing
Module Label C1.5

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Parallel and Distributed Systems

Learning Outcomes The goal of this course is to give the students a deep understanding of important techniques of
program analysis and program transformation. The emphasis lies on the acquiring of analytical
and technological competences: the students are enabled to analyse arbitrary programs by
applying the techniques of data and control dependency analysis and to perform optimizing
program transformation based on these analysis techniques. Examples are the vectorization
and parallelization of program parts or optimization towards a given memory hierarchy.
Methodical and algorithmis competences are acquired by learning scheduling and load
balancing algorithms and the underlying principles.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content The following topics are covered:

• Overview of current processor architectures and interconnection technologies

• Control flow and data flow analysis, data flow equations and solution methods for data
flow equations, optimizing transformations

• Data dependency analysis, loop dependencies, data dependence equations and solution
methods for them

• Program transformations for vectorization, parallelization and cache optimization

• Methods for scheduling and load balancing for instructions, loops, and tasks
• OpenMP programming
• Register allocation and program tranformations for reducing the register need of programs

• CPU programming with CUDA

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 240 h in total (90 h presence, 150 h preparation/post-processing with processing of worksheets)

Recommended C1.3: Parallel and Distributed Systems I

Prerequisites

Grading Written exam; For the admission to the written exam a vivid participation in the exercises is
required.

Frequency Every year in summer term

28
 Title Parallel Algorithms
Module Label C1.6

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Parallel and Distributed Systems

Learning Outcomes Students acquire in-depth knowledge about selected parallel algorithms from different fields of
application. In particular, in connection with exercises, students gain analytical and
methodological expertise, which empowers them to understand, implement, analyse, and design
parallel algorithms.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content Selected parallel algorithms are presented. The range extends from basic, widespread
algorithms (e.g., sorting) to complex algorithms from specific fields of application (e.g., computer
graphics). Emphasis is put on algorithms from the field of scientific computing. The exercises
cover theoretical problems as well as practical programming experience.

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 120 h in total (45 h presence, 60 h preparation/post-processing, 15 h preparation for exam)

Recommended Algorithms and Data Structures I, C1.3: Parallel and Distributed Systems I
Prerequisites

Grading Oral or written exam

Frequency Every year in summer term

29
 Title Programming and Data Analysis in Python
Module Label C1.7

Module Type Lecture

Area of Research Computer Science

Responsible Chair of Serious Games

Learning Outcomes Students learn to quickly prototype and implement numerical programs in Python. They learn
Python as a programming language and a scientific computing environment. They acquire
knowledge of the basic programming language, as well as of important libraries for scientific
computing, such as NumPy, SciPy, Matplotlib, Pandas, and TensorFlow/Keras. They develop
practical and applied skills in exploratory computing, rapid prototyping, and implementation of
numerical methods. In contrast to other environments, the Python scientific computing
environment is open source, widely used, optimized for programmer productivity, and benefits
from a large community and library ecosystem.

Content The Python programming language: Programming philosophy in Python, data types, control
structures, functions, object-oriented programming, debugging. Algorithms: Basic algorithms
(e.g., searching and sorting), bisection, recursion, dynamic programming, Newton’s method.
Matrix methods: Linear Algebra with NumPy, matrix factorizations, eigenvectors and values,
diagonalization, SVD, least squares and pseudoinverse. Data analysis: Pandas, clustering,
plotting. Neural networks and deep learning.

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (2 h/week)

Credit Points 4

Work Load 120 h in total (45 h presence, 60 h preparation/post-processing, 15 h preparation for exam)

Recommended None
Prerequisites

Grading Oral or written exam (85%), exercises (15%)

Frequency Every year in winter term

30
3.2 Module C2: Practical Course on Parallel Numerical Methods

Title Parallel Numerical Methods

Module Label C2

Module Type Pratical course

Area of Research Computer Science, Scientific Computing

Responsible Chairs of Parallel and Distributed Systems, Scientific Computing

Learning Outcomes • Implementation of parallel algorithms:

select suitable efficient numerical methods
choose data structures that are suitable for the respective problem
implement the numerical methods on a parallel computer using standard libraries

• Presentation and discussion:

choose and master suitable presentation techniques
speak freely about a subject and illustrate important structures instructively
answer spontaneous questions from the audience in a reliable manner

Content In this practical course, students implement manageable numerical problems (such as Gaussian
elimination, finite element discretization of 2d Laplacian, etc.) on parallel computers using the
programming language C/C++ and standard software libraries (LAPACK/BLAS, OpenMP,
OpenMPI). The resulting parallel efficiency is observed depending on the chosen
implementation (naive or advanced such as Schwarz methods).

Duration 1 semester

Language English

Teaching Method Practical course (2 weeks)

Credit Points 2

Work Load 50 h practical course; 10 h preparation for exam; in total: 60 h

Recommended A1: Numerical Methods for Differential Equations, C1.3: Parallel and Distributed Systems I,
Prerequisites D1.1: Efficient Treatment of Non-local Operators

Grading Implementation and presentation of approaches; active participation and discussion

Frequency Every year at the end of the winter term

31
4 Module Section D: Scientific Computing

4.1 Elective Modules D1: Complexity Reduction

Title Efficient Treatment of Non-local Operators

Module Label D1.1

Module Type Lecture

Area of Research Scientific Computing

Responsible Chair of Scientific Computing

Learning Outcomes • Understanding the way numerical algorithms for the solution of partial differential and
integral equations work

• Understanding that non-local operators may contain redundancies which can be used to
reduce their asymptotic complexity

• Ability to choose a suitable algorithm for a given class of partial differential and integral
equations

• Ability to implement the algorithms discussed in the lecture in a higher programming

language on a parallel computer

Content State-of-the-art linear complexity treatment of partial differential and integral operators and
parallelization techniques:

• fast multipole methods for the efficient treatment of multi-source potentials (one of the
TOP10 algorithms from the 20th century)

• hierarchical matrices (for the treatment of non-local operators with linear complexity)
• Schwarz methods (additive and multiplicative)

• Domain decomposition (overlapping and non-overlapping), BPS and Neumann-Neumann

preconditioners

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam; in total: 240 h

Recommended
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

32
 Title Fast Methods for Differential and Integral Equations
Module Label D1.2

Module Type Lecture

Area of Research Scientific Computing

Responsible Chair of Scientific Computing

Learning Outcomes • Understanding the way numerical algorithms for the solution of partial differential and
integral equations work

• Detection of suitable structures which can be exploited for the complexity reduction of
solution operators of elliptic boundary value problems

• Ability to choose a suitable algorithm for a given class of partial differential or integral
equations

• Ability to implement the algorithms discussed in the lecture in a higher programming

language

Content Optimal complexity solvers for the treatment of boundary value problems; efficient treatment of
parameter-dependent problems:

• subspace correction methods

• hierarchical bases and BPX preconditioners
• geometric and algebraic multigrid methods (convergence and implementation aspects)
• reduced bases methods

• analysis of hierarchical matrices

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam; in total: 240 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

33
 Title Efficient Numerical Treatment of Multiscale Problems
Module Label D1.3

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chair of Scientific Computing

Learning Outcomes

Content Methods for the efficient numerical treatment of multiscale problems such as MsFEM, HMM,
hp-FEM, GFEM/XFEM, and wavelet approximations will be presented.

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam; in total: 240 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

34
 Title Numerical Methods for Uncertainty Quantification
Module Label D1.4

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chair of Scientific Computing

Learning Outcomes

Content This is a cutting-edge area of Scientific Computing. In addition to Monte Carlo methods,
stochatic collocation, polyonomial chaos expansions, stochastic Galerkin methods, the
Karhunen-Loève expansion, model order reduction, and multilevel quadrature recent
developments in this area are to be discussed.

Duration 1 semester

Language English

Teaching Method Lectures (4 h/week) and tutorials (2 h/week)

Credit Points 4

Work Load 2 h lectures plus 3 h post-processing per week = 52,5 h;

1 h discussion sections plus 2,5 h preparation/post-processing = 52,5 h;
15 h preparation for exam; in total: 120 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

35
 Title High-dimensional Approximation
Module Label D1.5

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chairs of Applied and Numerical Analysis, Scientific Computing

Learning Outcomes By the end of the course, a successful student should

• understand the curse of dimensionality

• know several concepts to reduce the complexity in high-dimensional problems
• be able to apply such concepts to typical examples from finance, physics and engineering

Content • Introduction to problems from finance, physics and engineering leading to

high-dimensional partial differential equations, such as Black-Scholes and Fokker-Planck

• Modern concepts for high-dimensional problems including tensor product methods,

sparse grids, kernel-based methods, Monte-Carlo and Quasi-Monte-Carlo methods

• Error and stability analysis of such methods

• Efficient algorithms for and implementations of such methods

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1,5 h post-processing per week = 52,5 h;

1 h discussion sections plus 2,5 h preparation/post-processing = 52,5 h;
15 h preparation for exam; in total: 120 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

36
 Title Data Analytics
Module Label D1.6

Module Type Lecture

Area of Research Computer Science

Responsible Chair for Databases and Information Systems

Learning Outcomes Conceptual foundation of development of large databases (Big Data) and information systems
with focus on modeling. Deepening of proficiency in databases in the context of large and
complex database and web applications; imparting of interdisciplinary, analytical competences
for reconstructing and modeling complex applications (mostly stemming from the application
fields); technological competence for selecting and integrating heterogeneous modeling and
implementation concepts for the design and realization of data and process based applications.
Deepening of proficiency in the fields of data analytics. Realization of complex architectures in
the application fields Bio Informatics, Environmental Informatics and Engineer Informatics will be
discussed in all courses.
In contrast to students who have passed the corresponding bachelor module, students of this
modules can apply the previous techniques more autonomously and can relate them to formerly
acquired advanced skills.

Content first semester: Data Warehousing, Data Mining

second semester: Data Visualisation, Machine Learning, Ontologies, NoSQL, Distributed
Computing Concepts (MapReduce, Hadoop, etc.)

Duration 2 semesters

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 8

Work Load 4 h lectures plus 3 h post-processing per week = 105 h;

2 h discussion sections plus 5 h preparation/post-processing = 105 h;
30 h preparation for exam; in total: 240 h

Recommended Datenbanken und Informationssysteme I

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year during winter and summer term

37
 Title Complexity Reduction in Control
Module Label D1.7

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chair of Applied Mathematics

Learning Outcomes • Understanding why many numerical approaches to control problems suffer from the curse
of dimensionality

• Recognition of redundancies, such as the turnpike property in optimal control problems,

and their use for complexity reduction

• Knowledge about state-of-the-art model order reduction methods for linear and nonlinear
control systems and how they rely on the particular input-output-structure

• Ability to identify suitable techniques for particular applications

Content • State-of-the-art techniques for the reduction of complexity in control problems, as for
instance

– Model predictive control and its use as a complexity reduction technique

– Sparse grid and parallel computing approaches for Hamilton-Jacobi-Bellman
equations
– Model order reduction methods for control systems, such as balanced truncation or
proper orthogonal decomposition

• Implementation of selected algorithms in the tutorials

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended A2.3: Mathematical Control Theory, A1: Numerical Methods for Differential Equations
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

38
 Title Meshfree Methods
Module Label D1.8

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chair of Applied and Numerical Analysis

Learning Outcomes By the end of the course, a successful student should

• know fundamental techniques to reduce the complexity in meshfree discretisations for

solving partial differential equations numerically

• be able to choose a suitable meshfree method for a given problem

• master key methods of their analysis and implementation

Content • Short review of kernel-based collocation and particle methods for solving partial
differential equations

• Discussion of various cost reducing methods including:

– Fast summation techniques based on far field expansions of radial basis functions
(RBF) and completely monotone functions
– Local Lagrangian methods, RBF-FD
– Generalised alternating projection methods
– multilevel methods for compactly supported RBF
– Adaptive and greedy variants of the above methods

• Discussion of data structures and implementational details for meshfree methods

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam, in total: 120 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis,
Prerequisites A2.2: Constructive Approximation Methods

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

39
 Title Boundary Element Methods
Module Label D1.9

Module Type Lecture

Area of Research Numerical Mathematics

Responsible Chair of Scientific Computing

Learning Outcomes Exterior boundary value problems are difficult to treat via finite element discretizations due to the
unboundedness of the computational domain. This lecture presents a different approach by
which the boundary value problem is reformulated as an integral equation on the boundary. In
particular, this offers the advantage that only a lower-dimensional set has to be discretized. As a
consequence, the resulting linear systems are significantly smaller but fully populated in general.
The latter difficulty can be treated by local low-rank approximation.

By the end of the course, a successful student should

• know fundamental techniques to reduce suitable boundary value problems to boundary

integral equations

• master key methods of the analysis and implementation of boundary integral methods

• be able to reduce the complexity of suitable discrete non-local operators

Content • Sobolev spaces on manifolds

• fundamental solutions of partial differential operators
• boundary integral operators and their properties

• boundary integral equations and their finite element discretization

• generating the matrix coefficients
• fast boundary element methods

• applications: potential equation, linear elasticity, Stokes equations

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam, in total: 120 h

Recommended A1: Numerical Methods for Differential Equations, B1: Applied Functional Analysis
Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

40
 Title Optimization Methods in Machine Learning
Module Label D1.10

Module Type Lecture

Area of Research Optimization, Data Analysis

Responsible Chair of Applied Mathematics

Learning Outcomes • Knowledge about the special nature of optimization problems in machine learning
• Understanding of optimization algorithms in machine learning
• Ability to apply optimization algorithms in machine learning properly

Content • Introduction to machine learning

• Stochastic gradient methods
• Proximal gradient methods

• Acceleration techniques

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended Basic knowledge in numerics and optimization

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

41
4.2 Module D2: Special Skills in Scientific Computing

Title Special Skills in Scientific Computing

Module Label D2

Module Type Lecture

Area of Research Numerical Mathematics, Scientific Computing

Responsible Chairs of Applied and Numerical Analysis, Applied Mathematics, Scientific Computing

Learning Outcomes The goal of this module is to provide students with special skills in the areas of Numerical
Mathematics and Scientific Computing, relevant for current research activities.

Content An active field of mathematical research, in which specialized techniques are applied or in which
known techniques from different areas are combined in an original way. Examples are:

1. Fractional order differential operators

2. Optimization of nonsmooth problems

3. Advanced topics of optimization with partial differential equations

4. mathematical control theory for infinite-dimensional systems

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended Advanced status of study

Prerequisites

Grading Oral or written exam; active participation in the tutorials

Frequency Every year

42
 Title Efficient Algorithms for PDE Constrained Optimization
Module Label D2.1

Module Type Lecture

Area of Research Numerical Mathematics, Optimization

Responsible Chair of Applied Mathematics

Learning Outcomes
• ability to tackle complex and large scale PDE constrained optimization problems

• ability to identify and exploit inherent problem structure

• ability to construct and combine efficient solution techniques for PDE constrained
optimization

Content
• fast solvers and preconditioners of optimal complexity for linear problems
• efficient algorithms for nonlinear and constrained problems
• adaptive multi-level methods

Duration 1 semester

Language English

Teaching Method Lectures (2 h/week) and tutorials (1 h/week)

Credit Points 4

Work Load 2 h lectures plus 1.5 h post-processing per week = 52.5 h;

1 h discussion sections plus 2.5 h preparation/post-processing = 52.5 h;
15 h preparation for exam; in total: 120 h

Recommended B1: Applied Functional Analysis, A1: Numerical Methods for Differential Equations,
Prerequisites A2.5: Optimization of Differential Equations or A2.4: Nonlinear Optimization

Grading Oral or written exam; active participation in the tutorials

Frequency Every two years

43
5 Module Section E: Soft Skills

Title Soft Skills

Module Label E

Module Type Seminar

Area of Research

Responsible Programme coordinator

Learning Outcomes Key skills for personal development

Content Participation in seminars devoted to presentation skills, data processing, literature research,
handling of foreign-language literature, teamwork, etc.; the programme coordinator gives
individual recommendations

Duration 1 semester

Language English

Teaching Method Participation in seminars

Credit Points 2

Work Load In total 60 h of seminars

Recommended None
Prerequisites

Grading Proof of attendance

Frequency

44
6 Module Section F: Master’s Thesis

Title Master’s Thesis

Module Label F

Module Type Thesis

Area of Research all areas

Responsible Programme coordinator

Learning Outcomes Ability to prepare a scientific work (larger than a bachelor’s thesis)

Content Scientific work in the area of Scientific Computing that should have a connection with
application-driven questions and with the focus of this master’s programme. In particular,
interdisciplinary problems should be treated.

Duration 2 semesters

Language English or German

Teaching Method

Credit Points 30

Work Load 900 h (editing time: at most 10 months)

Recommended
Prerequisites

Grading Written thesis

Frequency

45
7 Recommended Curriculum
a) Full-time study

46
b) Part-time study

47