Today's technical systems - independent of their field of application - are generally characterized by a high level of complexity. Already the amount of requirements often makes it very difficult for a design engineer to keep pace. Classical, analytical methods are usually no longer sufficient for a comprehensive description and modeling of the system behaviour. Development and implementation often can only be realized with an extensive use of numerical simulation in combination with methods from the field of Computational Intelligence.
To solve complicated tasks in the development of complex systems, we focus on the combined usage of
- Computational Intelligence and Machine Learning to master high system complexity,
- Stochastic Simulation for a statistically relevant representation of requirements and data generation of numerical simulations and for robustness management,
- Software Process Models from software engineering for a requirement-driven process flow.
The applied methods are mostly oriented on natural approaches (>natural computation) and are very simple and robust at their core. Nevertheless (or just because of this), complicated real world problems can be solved by them in a very pragmatic way. This frees the resources to concentrate on the main engineering problem instead of struggling with the heavy mathematics of the solution techniques.
"If you have to solve quite complicated tasks, at least the solution methods should be easy to apply without building an additional layer of complexity on top."