A Machine Learning Approach to Support Procurement in Automotive Manufacturing
An approach to support decision making in product portfolio management by the example of E/E (electric-electronic) components is introduced. This is done from the procurement and development perspective of an automotive manufacturer. Due to the large number of variants of vehicles in the automotive industry, the electronic components to be integrated into the vehicle are not uniformly composed. These components are made available to the OEM (original equipment manufacturer) through external procurement instead of in-house production. The interplay of oligopolistic market conditions and the large number of different system configurations makes it difficult for OEMs to select the optimal supplier combination and the corresponding cost structures that will maximize profits for the company. To support the early component development stage the research outlined in this paper aims at the establishment of an evaluation model based on machine learning that is appropriate for electric E/E parts but is also transferrable to other raw material goods. After an introduction to theoretical basics and a description of the research framework, a machine learning model is presented that has been developed and tested. Preliminary results of this comparative study are shown.
Predictive Analytics; Machine Learning; Value Estimation; Procurement; Automotive Manufacturing.