Alexander Zienau
Research Subjects
Alexander Zienau has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
Eric Breitbarth
(4flow AG, Berlin, Germany
1 shared publication- Frieder Smolny1 shared publication
Guillaume Sagnol
1 shared publicationKarl Däubel
1 shared publicationLaura Gellert
1 shared publication- Martin Skutella
TU Berlin
1 shared publication
Latest Publications
- Open Access
- 23 May 2026
- Published by Elsevier in SSRN Electronic Journal
A Multiscale Optimization Framework for Real-World Transportation and Network Design Problems
The rapid growth of global supply chains and the increasing availability of data have madelogistics optimization a critical yet increasingly complex challenge for industries worldwide. Solving transportation and network design problems is essential for optimizing supply chains, yet the complexity of real-world instances often renders direct optimization infeasible. In addressing this challenge, various aggregation schemes have been proposed, attempting to exploit hidden global structure and redundancies in the input data. While practically successful, these methods do no allow for interactivity in the optimization or interpretation of a solution, and have been applied to models with linear or convex cost structures. We propose a novel framework that dynamically aggregates and refines network data, tariffs, and demands, enabling the efficient transfer of feasible solutions across scales. In contrast to traditional approaches we employ a dynamic data structure enabling independent modification of the level of detail in different parts of the network. We customize our framework to a multi-commodity flow model with piecewise linear costs and a facility location model with delivery time constraints, incorporating diverse industry-specific aspects validated by experienced logistics professionals. The method is evaluated on real-world data from the automotive industry, provided by the supply chain consulting company 4flow. Results demonstrate significantly improved solution quality and, beyond computational performance,highlight the practical value of the approach through its support for explainability and interactivity: decision-makers can navigate between different levels of aggregation to better understand, validate, and communicate optimization outcomes, facilitating adoption in real-world supply chain planning.
- 23 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Multiscale Optimization Framework for Real-World Transportation and Network Design Problems
The rapid growth of global supply chains and the increasing availability of data have madelogistics optimization a critical yet increasingly complex challenge for industries worldwide. Solving transportation and network design problems is essential for optimizing supply chains, yet the complexity of real-world instances often renders direct optimization infeasible. In addressing this challenge, various aggregation schemes have been proposed, attempting to exploit hidden global structure and redundancies in the input data. While practically successful, these methods do no allow for interactivity in the optimization or interpretation of a solution, and have been applied to models with linear or convex cost structures. We propose a novel framework that dynamically aggregates and refines network data, tariffs, and demands, enabling the efficient transfer of feasible solutions across scales. In contrast to traditional approaches we employ a dynamic data structure enabling independent modification of the level of detail in different parts of the network. We customize our framework to a multi-commodity flow model with piecewise linear costs and a facility location model with delivery time constraints, incorporating diverse industry-specific aspects validated by experienced logistics professionals. The method is evaluated on real-world data from the automotive industry, provided by the supply chain consulting company 4flow. Results demonstrate significantly improved solution quality and, beyond computational performance,highlight the practical value of the approach through its support for explainability and interactivity: decision-makers can navigate between different levels of aggregation to better understand, validate, and communicate optimization outcomes, facilitating adoption in real-world supply chain planning.
- 18 February 2021
- Published by Emerald Publishing in Journal of Humanitarian Logistics and Supply Chain Management
- Vol. 11 (2) , 227-247
Protecting vulnerable people during pandemics through home delivery of essential supplies: a distribution logistics model
Purpose: This paper studies a concept for protecting vulnerable population groups during pandemics using direct home deliveries of essential supplies, from a distribution logistics perspective. The purpose of this paper is to evaluate feasible and resource-efficient home delivery strategies, including collaboration between retailers and logistics service providers based on a practical application.Design/methodology/approach: A food home delivery concept in urban areas during pandemics is mathematically modeled. All seniors living in a district of Berlin, Germany, represent the vulnerable population supplied by a grocery distribution center. A capacitated vehicle routing problem (CVRP) is developed in combination with a k-means clustering algorithm. To manage this large-scale problem efficiently, mixed-integer programming (MIP) is used. The impact of collaboration and additional delivery scenarios is examined with a sensitivity analysis.Findings: Roughly 45 medically vulnerable persons can be served by one delivery vehicle in the baseline scenario. Operational measures allow a drastic decrease in required resources by reducing service quality. In this way, home delivery for the vulnerable population of Berlin can be achieved. This requires collaboration between grocery and parcel services and public authorities as well as overcoming accompanying challenges.Originality/value: Developing a home delivery concept for providing essential goods to urban vulnerable groups during pandemics creates a special value. Setting a large-scale CVRP with variable fleet size in combination with a clustering algorithm contributes to the originality.
- 18 February 2021
- Published by Emerald Publishing in Journal of Humanitarian Logistics and Supply Chain Management
- Vol. 11 (2) , 227-247
Protecting vulnerable people during pandemics through home delivery of essential supplies: a distribution logistics model
Purpose: This paper studies a concept for protecting vulnerable population groups during pandemics using direct home deliveries of essential supplies, from a distribution logistics perspective. The purpose of this paper is to evaluate feasible and resource-efficient home delivery strategies, including collaboration between retailers and logistics service providers based on a practical application.Design/methodology/approach: A food home delivery concept in urban areas during pandemics is mathematically modeled. All seniors living in a district of Berlin, Germany, represent the vulnerable population supplied by a grocery distribution center. A capacitated vehicle routing problem (CVRP) is developed in combination with a k-means clustering algorithm. To manage this large-scale problem efficiently, mixed-integer programming (MIP) is used. The impact of collaboration and additional delivery scenarios is examined with a sensitivity analysis.Findings: Roughly 45 medically vulnerable persons can be served by one delivery vehicle in the baseline scenario. Operational measures allow a drastic decrease in required resources by reducing service quality. In this way, home delivery for the vulnerable population of Berlin can be achieved. This requires collaboration between grocery and parcel services and public authorities as well as overcoming accompanying challenges.Originality/value: Developing a home delivery concept for providing essential goods to urban vulnerable groups during pandemics creates a special value. Setting a large-scale CVRP with variable fleet size in combination with a clustering algorithm contributes to the originality.