A spare part for a faulty refrigerator? Every end customer always wants to get their hands on one soon as possible. This poses a number of challenges for producers. On the one hand, companies want to reduce the costs of storage and transport as well as the capital commitment. On the other hand, the service quality of a manufacturer is also measured by how quickly they can deliver spare parts. To resolve this dilemma, BSH Hausgeräte is cooperating with the ADA Lovelace Centre.
The ADA Lovelace Center for Analytics, Data and Applications in Nuremberg was founded at the end of 2018 by the Fraunhofer Institute for Integrated Circuits IIS in cooperation with the Friedrich-Alexander-University Erlangen-Nuremberg and Ludwig-Maximilians-University Munich. It combines research on artificial intelligence with industrial applications in a special way. The Center brings together local, regional and national players and cooperates with international partners such as the Machine Learning Center at the Georgia Institute of Technology and the RIKEN Center for Advanced Intelligence in Tokyo.
Optimisation of spare parts logistics
BSH Hausgeräte is the largest manufacturer of household appliances in Europe, producing brands such as Bosch, Siemens, Gaggenau and Neff. The breadth of the portfolio was also one of the challenges, as many brands means many spare parts. Together, they developed an AI application that now helps the company predict the need for spare parts. ‘With this tool for long-term prognosis, various methods of machine learning are used to enable reliable predictions’, explains Institute Director Prof. Alexander Martin. The tool has now been implemented and is helping BSH Hausgeräte to reduce its inventory to a level that is actually necessary.
‘Many companies do not know which data they have available and in what quality and quantity, which data they need to solve a specific application or vice versa and which applications they can actually optimise using their data’, explains the business mathematician. Research at the ADA Lovelace Center helps to set the course by supporting companies in making these fundamental decisions. ‘Our contribution consists of methodological skills in data analysis, whereby we cover the entire range: from classical description, forecast and prognosis to decision-based methods’, says Martin. For this purpose, the Center currently relies on eight areas of expertise, such as automatic, adaptive learning, mathematical optimisation methods or semantics, which are being further developed in 10 application fields, including in the areas of data-driven localisation, self-optimising, adaptive logistic networks and driver assistance systems in rail traffic.
The decisive factor here is the proximity to practice. ‘The special thing about it is that we further develop the competencies, methods and procedures using the example of applications. This means that we work on use cases that come from industrial practice’. Martin and his team have even developed a new cooperation format for this collaboration: joint labs. Here, scientists and company employees work as part of small, agile, interdisciplinary development teams on company-specific issues for a limited period of time. ‘This innovative form of collaboration and infrastructure allows more creative solutions to be developed in less time’, says Prof. Martin.
Shorter distances in warehouse logistics
AI offers many possible solutions, especially in logistics, for example in the area of warehouse logistics: here, the methodological know-how of the ADA Lovelace Center was used to develop optimisation software that supports the international logistics service provider Schnellecke in its warehouse processes. How can products be stored in a high-bay warehouse in such a way that the most efficient retrieval is possible? This is a far from simple question as all sorts of conditions, from the specified loading and unloading sequence to fire protection, had to be taken into account. ‘Our experts in the field of application used so-called Mixed Integer Programming – a mathematical optimisation method. This resulted in the development of software that enables goods to be stored in such a way that the distances travelled by the order pickers are significantly shorter’, says Martin.
Opportunities for logistics
‘AI in itself is not a new topic, but is currently experiencing a great deal of hype due to e.g. improved computer and software performance’, explains Institute director Martin. And there are also opportunities for industry and especially for logistics as a cross-sectional industry with a very dynamic environment: ‘This allows problems of very high complexity and data size to be solved in real time – true treasures for companies are hidden here that need to be unlocked’. The huge potential of AI should not make you blind to your business, however, warns Martin. ‘We must not make the mistake of thinking of AI as the sole all-purpose weapon, but rather, depending on the issue at hand, we need to understand it as an important tool within an overall context’.
- Artificial intelligence & robotics