Research

Areas of research and publications

The chair’s research lies at the interface of mathematical optimization, stochastic programming, and application-oriented decision support in production, supply chain, and transportation. Methodologically, key areas are mixed-integer and stochastic optimization, decomposition and cutting-plane methods, and the coupling of optimization and simulation. Regarding applications, robust and risk-averse planning approaches for automotive and pharmaceutical production prevail, as do resilient, partly “green” transport and distribution problems.

Our aim is to connect theoretical and methodological progress with industrial relevance. Our algorithmic contributions to date (dual‑simplex implementations, cutting planes for mixed‑integer optimization, and variants of Benders decomposition) accelerate the solution of real‑size stochastic planning models. In the application domains, we demonstrate that integrated, scenario‑based optimization—partly combined with simulation—enables better decisions than classical deterministic models and sequential planning: fewer sequence violations and bottlenecks in automotive production, lower shortage risk in pharmaceutical networks, and more robust as well as climate‑aware transport plans.

 

Publications (PDF)

Google Scholar

Computational algorithms and stochastic optimization

In this focus area, we work on the efficient solution of large optimization problems—from fast LP solvers to advanced decomposition and cutting‑plane approaches for stochastic programs. Our contributions aim to make stochastic and risk‑averse models practically usable—through fewer iterations, stronger relaxations, and robust runtimes at realistic scales. They provide the methodological foundation for the application‑oriented results in the other two research foci. The following results have been achieved in this focus area to date:

Dual simplex and cuts: With work on stable, fast implementation techniques of the dual simplex [1,2,3] and on pivot‑and‑reduce cuts to strengthen mixed‑integer Gomory cuts [4], we have advanced the practical performance of standard MIP/LP solvers.

Benders/L‑shaped methods and cut management: Contributions to the nested L‑shaped method with dynamic sequencing and cut consolidation [6] and to the use of “oracles of on‑demand accuracy” [7] address the core challenge in two‑stage stochastic models: generating few but effective cutting planes and reducing runtime. A more recent result is the combination of L‑shaped with strengthened lift‑and‑project cutting planes [9], which accelerates convergence through stronger master relaxations.

Risk optimization in decomposition methods: In [8], we demonstrate how risk‑averse objectives (e.g., CVaR) can be incorporated into two‑stage stochastic optimization models and develop new algorithmic techniques (e.g., scenario bundling, cut‑based tuning) to improve tractability.

Software and modeling: With the solver system PNBSolver [5,6], we address the gap between modeling convenience and powerful solution of stochastic programs. It enables the implementation of stochastic programs in the FlopC++ modeling language and their solution via efficient, Benders‑based decomposition procedures.

Current and future research: We are currently working on further improving the developed methods by integrating machine learning and artificial intelligence techniques.

Selected publications:

  1. Koberstein. The Dual Simplex Method: Techniques for a fast and stable implementation. Dissertation at the University of Paderborn, Fakultät für Wirtschaftswissenschaften / Department Wirtschaftsinformatik, 2005. http://digital.ub.uni-paderborn.de/hsmig/content/titleinfo/3885

  2. A. Koberstein and U. H. Suhl (2007). Progress in the Dual Simplex Algorithm for Solving Large Scale LP Problems: Practical Dual Phase 1 Algorithms. Computational Optimization and Applications 37 (1): 49-65.

  3. A. Koberstein (2008). Progress in the Dual Simplex Algorithm for Solving Large Scale LP Problems: Techniques for a fast and stable implementation. Computational Optimization and Applications 41 (2): 185-204. 

  4. F. Wesselmann, A. Koberstein, U. Suhl (2011). Pivot-and-Reduce Cuts: An Approach for Improving Gomory Mixed-Integer Cuts. European Journal of Operational Research 214 (1): 15-26.

  5. C. Wolf, A. Koberstein, T. Hultberg (2011). Stochastic Extensions to FlopC++. In: B. Hu et al. (eds.) Operations Research Proceedings 2010. Springer-Verlag Berlin Heidelberg. 2011. pp. 333-336.

  6. C. Wolf, A. Koberstein (2013). Dynamic sequencing and cut consolidation for the parallel hybrid-cut nested L-shaped method. European Journal of Operational Research 230:143-156.

  7. Wolf, C. Fábián, A. Koberstein, L. Suhl (2014). Applying oracles of on-demand accuracy in two-stage stochastic programming – a computational study. European Journal of Operational Research 239, 437-448.

  8. C. Fábián, C. Wolf, A. Koberstein, L. Suhl (2015). Risk-averse optimization in two-stage stochastic models: computational aspects and a study. SIAM Journal on Optimization 25(1):28-52.

  9. Glushko, P., Fábián, C. I., Koberstein, A. (2022). An L-shaped method with strengthened lift-and-project cuts. Computational Management Science 19, 539–565. https://doi.org/10.1007/s10287-022-00426-y

Production and supply chain planning under uncertainty (in automotive, pharma, global networks)

A second, extensive field comprises integrated models for production, capacity, and network decisions under demand, process, and regulatory uncertainty—with constant attention to implementability in industry.

Planning in the automotive industry: Ranging from strategic‑tactical planning models for global network planning under uncertainty [1] via models for aggregate production planning with workforce flexibility [4] to short‑term production and sequencing procedures, we investigate an end‑to‑end chain of planning problems in the automotive industry. Work on reconfiguration [2] and resequencing [8,9] aims to provide practice‑oriented approaches for effectively coordinating mixed‑model flow lines. In more recent studies on anticipating technical car‑sequencing rules already at the master production scheduling stage [13,15], we show that early, integrated consideration of downstream sequencing constraints can reduce overloads and sequencing rule violations in final assembly under logistical disruptions. Work on robust delivery profiles in area‑forwarding networks combines cost and ecological objectives [3,7] and shows that stochastic profile selection can simultaneously reduce emissions and delay risks in supply procurement and inbound logistics.

Postponement and flexibility in supply chains: In several contributions we develop and evaluate stochastic models for decoupling points and delayed product differentiation [5,10]. We provide a structured literature review in [11]. As a result, we show that the principle of postponement can increase service levels and reduce stockouts, especially under volatile, hard‑to‑forecast demand—and identify thresholds (volatility, product variety, changeover times) beyond which these strategies are economical.

Pharmaceutical networks and regulatory uncertainty: Models for network planning under uncertain production approvals [12] and for avoiding supply shortages through postponement and flexibility [14] quantify how regulatory lead‑time risks can be incorporated into capacity and inventory decisions. Key findings: Flexible formulation and packaging stages and late product assignment reduce the risk of stockouts and lower total inventories at given service levels.

Global production networks and financial hedging: An integrated model couples location/capacity decisions with currency hedging [6]. Result: Financial hedges are not an “add‑on” but change optimal network structures; without integrated consideration, both risk and expectation are systematically misestimated.

Current and future research: We are currently interested in modeling operational planning problems as sequential decision processes and in applying reinforcement learning methods.

Selected publications:

  1. R. Bihlmaier, A. Koberstein, R. Obst. Modeling and optimizing of strategic and tactical production planning in the automotive industry under uncertainty (2009). Operations Research Spectrum 31: 311-336.

  2. S. Altemeier, M. Helmdach, A. Koberstein, W. Dangelmaier (2010). Reconfiguration of Assembly Lines under the Influence of High Product Variety in the Automotive Industry – A Decision Support System. International Journal of Production Research 48 (21): 6235-6256.

  3. T. Schöneberg, A. Koberstein, L. Suhl (2010). An optimization model for automated selection of economic and ecologic delivery profiles in area forwarding based inbound logistics networks. Flexible Services and Manufacturing Journal 22 (3-4): 214-235.

  4. T. Sillekens, A. Koberstein, L. Suhl (2011). Aggregate production planning in the automotive industry under special consideration of workforce flexibility. International Journal of Production Research 49 (17): 5055-5078.

  5. S. Guericke, A. Koberstein, F. Schwartz, S. Voß (2012). A Stochastic Model for the Implementation of Postponement Strategies in Global Distribution Networks. Decision Support Systems 53 (2): 294–305.

  6. A. Koberstein, E. Lukas, M. Naumann (2013). Integrated Strategic Planning of Global Production Networks and Financial Hedging under Uncertain Demand and Exchange Rates. Business Research (BuR) 6 (2), 215-240.

  7. T. Schöneberg, A. Koberstein, L. Suhl (2013). A Stochastic Programming Approach to Determine Robust Delivery Profiles in Area Forwarding Inbound Logistics Networks. Operations Research Spectrum 35 (4), 807-834.

  8. C. Franz, E.C. Hällgren, A. Koberstein (2014). Resequencing orders on mixed-model assembly lines: Heuristic approaches to minimise the number of overload situations. International Journal of Production Research 52 (19), 5823-5840.

  9. C. Franz, A. Koberstein, L. Suhl (2015). Dynamic resequencing at mixed-model assembly lines. International Journal of Production Research 53 (11).

  10. C. Weskamp, A. Koberstein, F. Schwartz, L. Suhl, S. Voß (2019). A two-stage stochastic programming approach for identifying optimal postponement strategies in supply chains with uncertain demand. Omega 83:123–138. https://doi.org/10.1016/j.omega.2018.02.008

  11. Blossey, G., Hahn, G. J., & Koberstein, A. (2021). Managing Uncertainty in Pharmaceutical Supply Chains: A Structured Review. In Proceedings of the 54th Hawaii International Conference on System Sciences (p. 1435).

  12. G. Blossey, G. J. Hahn, A. Koberstein (2022). Planning pharmaceutical manufacturing networks in the light of uncertain production approval times. International Journal of Production Economics 244:108343. https://doi.org/10.1016/j.ijpe.2021.108343

  13. T. Krüger, A. Koberstein, N. Bittner (2022). Anticipating technical car sequencing rules in the master production scheduling of mixed-model assembly lines. Flexible Services and Manufacturing Journal 34, 351–407. https://doi.org/10.1007/s10696-021-09443-6

  14. Blossey, G., Hahn, G. J., Koberstein, A. (2025). Preventing drug shortages through improved demand fulfillment: The untapped potential of postponement and flexibility, International Journal of Production Economics, 109902, https://doi.org/10.1016/j.ijpe.2025.109902

  15. Krueger, T., Koberstein, A. Plant-wide master production scheduling in the automotive industry under component blockings: an MILP-approach and a simulation study. Journal of Business Economics (2026). DOI: 10.1007/s11573-026-01264-z

Transport, routing, and logistics optimization for resilience and sustainability

The third focus addresses modern transport problems—from green fleets and energy constraints to humanitarian logistics—often under uncertainty. The research provides practical models for more resilient and lower‑emission transport networks. It highlights how technological options (hybrid drives, drones, “green” fleets) interact with network‑design and deployment decisions and how to configure them optimally under uncertainty.

Energy‑ and emissions‑aware routing: With the Hybrid‑Electric‑Vehicle Traveling Salesman Problem [1] and its extension with time windows [4], energy states, recuperation, and charging decisions are integrated into routing. Result: New formulations and algorithms show when intermediate charging and adjusted speeds ensure feasibility and reduce costs; classical TSP heuristics are insufficient under energy/time constraints. The hub‑location problem with a mixed green fleet [8] quantifies emission–cost trade‑offs and shows how fleet mix and node structure should be jointly chosen.

Maritime and humanitarian logistics under uncertainty: For the stochastic repositioning of liner shipping fleets [5] and in a comprehensive overview of uncertainty in maritime routing and scheduling [6], demand and travel‑time uncertainties are modeled and mitigation strategies (buffers, flexible rotations) derived. A stochastic model for drones in earthquake relief [7] shows that appropriate scenario policies significantly reduce response times and failure risks—especially under highly volatile demand and infrastructure disruptions.

Storage and terminal systems: Work on puzzle‑based storage systems [2] with an application to grid‑based early‑baggage storage [3] develops optimal/heuristic algorithms for layout, movement, and retrieval. Simulations demonstrate how movement rules and the number of escorts affect throughput and congestion dynamics.

Current and future research: Here, we apply newer distributionally robust, data‑driven optimization approaches to provide planners with more realistic cost assessments and robust plans under limited historical data.

Selected publications:

  1. C. Doppstadt, A. Koberstein, D. Vigo (2016). The Hybrid Electric Vehicle - Traveling Salesman Problem European. European Journal of Operational Research 253, 825-842.

  2. A. Yalcin, A. Koberstein, K.-O. Schocke (2019). An optimal and a heuristic algorithm for the single-item retrieval problem in puzzle-based storage systems with multiple escorts, International Journal of Production Research 57(1), 143-165. DOI: 10.1080/00207543.2018.1461952

  3. A. Yalcin, A. Koberstein, K.-O. Schocke (2019). Motion and layout planning in a grid-based early baggage storage system: Heuristic algorithms and a simulation study. OR Spectrum 41, 683–725. https://doi.org/10.1007/s00291-018-0545-z

  4. C. Doppstadt, A. Koberstein, D. Vigo (2020). The Hybrid Electric Vehicle-Travelling Salesman Problem with time windows. European Journal of Operational Research 284(2):675–692. https://doi.org/10.1016/j.ejor.2019.12.031

  5. S. Kuhlemann, J. Ksciuk, K. Tierney, A. Koberstein (2021). The stochastic liner shipping fleet repositioning problem with uncertain container demands and travel times. EURO Journal on Transportation and Logistics 10:100052. DOI: 10.1016/j.ejtl.2021.100052

  6. J. Ksciuk, S. Kuhlemann, K. Tierney, A. Koberstein (2023). Uncertainty in maritime ship routing and scheduling: A Literature review. European Journal of Operational Research, 308(2), 499–524. DOI: 10.1016/j.ejor.2022.08.006

  7. Dukkanci, O., Koberstein, A., & Kara, B. Y. (2023). Drones for relief logistics under uncertainty after an earthquake. European Journal of Operational Research, 310(1), 117–132. DOI: 10.1016/j.ejor.2023.02.038

  8. Dukkanci, O., Campbell, J. F., Koberstein, A. (2025). Hub location problem with a mixed green fleet. European Journal of Operational Research, 330(1), 84–99. DOI: 10.1016/j.ejor.2025.08.033

Prof. Dr. Achim Koberstein