The Simulation and Optimization group offers software practicals for bachelor and master students (or equivalent, e.g., Diplom) who aim for a degree in mathematics, physics, or computer science.

Software practicals are not seminars, but are project oriented. Students work alone or in small groups on a well-defined project under the supervision of one or two members of the group (PhDs, PostDocs, advanced MSc/Diplom students). The number of students allowed per practical depends on the project. Meetings with the supervisors are scheduled personally and on demand.

The practicals are divided into beginners' and advanced practicals. The workload of a beginners' practical is 40 hours per person. The workload of an advanced practical is 80 hours per person. Beginners' practicals also have lower requirements on the students' skills.

Open Software Practicals

• the names of the students in the group,
• their respective aspired degree (e.g., MSc in mathematics with minor computer science),
• relevant courses taken (e.g., Informatik I, Numerik 0-2, etc. or similar for non-Heidelberg students),
• a concise comment on the personal motivation for undertaking the selected practical and on prospects for continuing with a Diplom, MSc, BSc, or even a PhD degree in the Simulation and Optimization group.

Advanced

• Solving Symmetric Indirect Systems on GPU.
• Testing, Extending and Documenting Python Interface to Integrator SolvIND.
• Fast convergence for trust region solver.
• Python Interface to Benchmark Library MacMPEC.
• Interfacing Integrator SolvIND with Automatic Differentiation Packages CasADi, CppAd and Adept.
• Factorization based linear algebra for trust region solver.
• Parallel Optimal Control of the 3D Heat Equation.

Beginners

Running software practicals

• Pseudospectral Optimal Control.
  (Sankari Ganesan, Jing Xu)
• Preparation of aerospace test cases for optimization with MUSCOD.
  (Darshan Balaganchi Muralidhara, Bryan Wolfford)
• Shooting Methods in Practice: Control of a USB Missile Launcher.
  (Hannes Koderisch, Andreas Neudecker)
• Data Preparation of Stellar Atmosphere Simulation Results.
  (Roger Gaczkowski)

Examples of completed software practicals

• 2D Slot Car Visualization.
• Rankine-Cycle Visualizations.
• Parameter estimation for stochastic models in Systems Biology.
• Evaluation of an algebraic multi-grid method
  (Ruben Garske)
• Levenberg-Marquardt algorithm for parameter estimation.
  (JRG Optimum Experimental Design)
• Implementation of rigorous tests for a linear algebra software package.
  (Artjom Zern)
• Metaprogramming for the formulation of complex optimization problems
  (Christian Amrhein, Tristan Engel, Lisa Kolb)
• Local Minima in Parameter Estimation Problems for a Predator Prey Model.
  (Shashank Jaiswal)
• Simulation and Visualization of Robot Kinematics
  (Ijaz Ahmad, Philipp Stekl)
• Automatic Differentiation of Optimal Control Problem Functions
  (Shashank Jaiswal)
• Initial Satellite Orbit Determination with Nonsingular Kepler Elements
  (Jamin Hass)
• Design and Implementation of a new parser for MUSCOD-II input files using the boost::spirit parser.
  (Timo Minartz, Markus Breit)
• Implementation of a new visualization frontend for MUSCOD-II/SOLVIND using GTK+ and Cairo
  (Timo Taglieber)
• SolvIND/DAESOL-II Integrator Benchmark and Test
  (Manuela Ernst)
• Visualization of Flight Maneuvers
• C++ Implementation of a Nonlinear Kalman Smoother for Marker Data Processing
• Setting up a complex evaporator model with MUSCOD
• GUI for DistMesh
• MUSCOD Tutorial
• Java Programming in a Literature Database Project
• Portable MUSCOD
• SONIC++: A C++ implementation of matrix and vector template classes on top of BLAS
(Michael Hoffer, Christian Poliwoda)

Contact

For general questions please contact