Software Practical: 

Algorithmic Generation of Sensitivities ODEs with Implicitly Switched Right Hand Side

The practical deals with the algorithmic generation of sensitivities in ordinary differential equations with implicit (state-dependent) nondifferentiability points in the right-hand side, given as Matlab program code with functions such as min, max, abs, sign, as well as if-branches. The current software determines non-differentiability points up to machine precision, switching functions are automatically generated and outputs them as Matlab program code.

Currently, the software utilizes Matlab's ode45 numerical ODE solver. In the practical, the software shall be adjusted to being able to use also the other Matlab-Solvers, especially ode15s. For that, a programmatic framework is to be determined for easy switching between different Matlab solvers.

Focus

• Learning how to correctly treat ODEs with non-differentiable right hand side functions
• Learning and using internals of Matlab for source code processing and generations

Classification of the project

This project is suited for one student as a beginner's practical, but can be extended to an advanced software practical. The practical can serve as a preparation project for a Bachelor or Master thesis in the Simulation and Optimization or in the Numerical Optimization Group (Prof. Kostina) group.

The project is to be completed by extending existing software code and a report and an oral presentation in the group's seminar.

Contact:   andreas.sommer@iwr.uni-heidelberg.de