Master thesis: Machine learning for job scheduling in matrix production (international students are very welcome)

17. Februar 2020

Production systems based on the matrix principle are regarded as a promising approach to deal with the traditional conflict of objectives between productivity and flexibility. Against the background of increasing variant diversity, such systems are currently being discussed extensively in science and practice, especially in the assembly area, in order to ensure the competitiveness of manufacturing companies in high-wage locations.

In this thesis, methods from the field of machine learning (ML) are used to estimate completion times. For this purpose, it is first necessary to qualitatively compare other existing approaches such as queuing models. Subsequently, an anylogic simulation model of a matrix production will be used to generate data for the ML approach. The simulation model already exists at our chair, but has to be adapted and extended for this work. Therefore solid knowledge/previous experience with JAVA (Anylogic programming language) is a necessary prerequisite for this work. The actual ML part of the work is to be done with Python. In addition to the setup of the simulation and ML environment, the work should ideally answer the question of which factors significantly influence the estimation quality and how well completion times can be estimated in advance using approaches of machine learning in matrix production.

Start: immediately (as of 17.02.2020)

The LFO offers:
Exciting topics in the topic area Industry 4.0 & autonomous production systems
– High learning effect through close supervision
– Possibility for phD & contacts to well-known industrial companies
– Flexible time management and nice & familiar working environment (regular joint after work events like playing pool)

What I desire:
– Degree in (business) informatics, statistics or business engineering/mechanical engineering/logistics with previous IT experience
– Affinity for Production & Logistics 4.0
– Previous knowledge through internships/work student activities/ project work/ study modules etc.
– Self-confidence, initiative and intrinsic motivation
– Sufficient English language skills (work will be done in english)
– JAVA skills are necessary, Python and Anylogic experiences would be helpful
– Duration: approximately 6 months

Contact:
Daniel Mueller
mueller@lfo.tu-dortmund.de
0231 / 755 – 7326
I look forward to your application.