The production facilities of modern enterprises are subject to a trend towards higher complexity and linking-up. This trend is intensified by the dynamics created in the facility structure through frequent product changes and shortened product life cycles. The resulting development poses a significant challenge for the personnel on all levels.

One of the measures to master the situation consists of an increased acquisition and evaluation of production- and facilities data. Their technical or economical analysis and interpretation with respect to operative or planning decisions on the one hand is difficult while on the other hand it is known that in the production area the required data mining tools are not widespread: due to a poll performed by KDnuggets for example less than 10% of the polled enterprises regularly use data mining in the production area. That is where the project ProDaMi joins in, focussing on the fields of process control and production planning and - control.

Goals

ProDaMi is supposed to enable the project partners to efficiently develop tailor-made data mining solutions for the production sector. To this end and on the basis of five demonstrating applications the following topics are dealt with:

• Development resp. collection of a library of relevant, particularly suited data mining routines including specifications for their parametrization.
• Development resp. buildup of a collection of routines for data preprocessing and import of common formats.

The usage of standardized interfaces and an internal data format grants the compatibility of all routines among each other. Parametrization of complex mining methods is supposed to be performed largely automatic, data-based (if not otherwise desired). Implementations of concrete applications can be provided in different languages. Currently C++, Matlab and VB/VBA are scheduled.

 

Offering

The department SYS offers services in the area "analysis and optimization of continous, complex production processes":

• development of tailor-made data mining tools e.g. to increase quality,
  and efficiency, and for the design of experiments,
• execution of data analyses,
• consulting in statistics and data dining.