PrEvelOp

Initial situation

The challenge for manufacturing companies is to control the variance in their production programs in order to survive in a competitive environment with higher cost pressure. In addition to discontinuations, standardization measures are one of the main levers for reducing production program variance and increasing profitability.

Solution approach

Within the project, functions are developed that identify similarities in order data (consisting of item as well as manufacturing process data) using unsupervised learning methods. The functions enable the grouping of orders according to article-specific and/or process-specific characteristics. Depending on the groupings, suitable measures for reducing the manufacturing program variance are automatically suggested by decision support.

Expected result

As a result, an open source library is targeted - with functions for ML-based reduction of variance in the manufacturing program, which will be made available to SMEs via a permissive license and also for testing and practice purposes in the form of a demo application. In addition, the project aims to promote the exchange of best practices for mastering item and manufacturing process variance between the project partners.

Benefits for the target group

Minimizing manufacturing program variance is a necessary prerequisite for identifying and implementing variance-reducing measures in order to address variance-induced complexity cost drivers - such as small batch sizes or frequent tool changes - and to reduce the dilemma between individual production (‘economies of scope’) and mass production (‘economies of scale’).