Fuzzy Rule-Based Machine Learning in Process Modeling

The optimization of technical processes requires a sufficient understanding of the interrelations in a process. More specifically, this means that a model of the given process is needed. Traditional methods for modeling technical processes are based on analytical relationships that are either explicit formulas derived from knowledge in physics, chemistry or engineering science or given as solutions of systems of differential equations. In practice, however, one is often confronted with processes that are too complex to be sufficiently captured by an analytical model or — although it would be possible in principle — the development of a full-fledged analytical model is too costly.