TU Berlin and 4flow develop multiscale data structures and algorithms

Simplifying complex supply chain networks the smart way

How can very large networks be designed more intelligently and efficiently? 4flow collaborated with the Combinatorial Optimization & Graph Algorithms Group (COGA) at the Technical University of Berlin in Germany in order to answer this question.

Today’s supply chain networks are highly complex and increasingly interconnected around the globe. Taken together, the data available in such networks often exceeds the capacity of existing planning processes and consequently hinders computer-assisted optimization. If this data is irreversibly simplified – as has been standard practice until now – or if subnetworks are optimized separately, the solutions derived from operational planning often cannot be implemented in the original network.


Visualization of a complex supply chain network

In the LogiScale project, COGA and 4flow conceptualized and developed a multiscale representation of supply chain networks. This representation facilitates the visualization of cost, time, and location data in various levels of aggregation (scales). “This allows for the aggregation of detailed data and solutions on smaller scales, coupled with broader solutions on larger scales, all while preserving the input data – which can then be used to derive optimization algorithms,” said Dr. Martin Skutella, Einstein professor for mathematics and computer science at the Technical University of Berlin. The multiscale representation allows users to examine data at several levels of granularity simultaneously and continuously.

“While it may sound complicated, it is extremely useful in supply chain practice. From to-the-hour details for tactical network planning which considers all modes of transportation, to cost and service level optimization in distribution networks, all the way to improved packing algorithms, the concrete applications developed during the project enable this kind of multiscale representation,” said Dr. Laura Gellert, who led 4flow’s involvement in the project. For this reason, the 4flow researchers were joined by consultants and supply chain planners in the project. Currently, the software developers are implementing the project results in 4flow vista®, the integrated standard software for supply chain design and transportation planning.

After three and a half years of excellent real-world results, the project has come to a close. The project partners will continue their successful collaboration with the support of the Berlin University Alliance cluster of excellence MATH+. 4flow is simultaneously working on many other innovation projects, the foremost of which is identifying use cases for artificial intelligence in the optimization and management of complex international supply chains.

LogiScale was funded by the European Union’s Fund for Regional Development and the Investitionsbank Berlin (IBB).