The OMDS/BALOR Talk Series invites you to the following seminar talks in Winter Semester 2025/2026:
- Speaker: Saverio Basso (Dalle Molle Institute for Artificial Intelligence, University of Applied Sciences and Arts of Southern Switzerland)
- Title: Learning Dominant States in Elementary Resource Constrained Shortest Path Problems
- Time: Wednesday, Jan. 21, 2026, 16:30 - 17:30
- Location: OMDS Seminar Room (03.307, Oskar-Morgenstern-Platz 1)
- Abstract: Elementary Resource Constrained Shortest Path Problems (ERCSPPs) are central to optimization and transportation research. They involve finding the shortest path under resource constraints while ensuring that nodes are not revisited. The problem is strongly NP-hard and frequently arises in orienteering tasks as well as in subproblems of Vehicle Routing and Crew Scheduling. In this study, we explore whether machine learning can be used to identify promising states within dynamic programming algorithms, commonly applied to solve ERCSPPs, with the ultimate goal to improve search decisions. We start by solving 41 single-resource instances from SPPRCLIB using iterative relaxation approaches via the PathWyse library, collecting constant-time features for all generated dynamic programming states. These features allow us to build two large datasets, totaling several hundred million labels. We then apply machine learning to analyze them, uncovering patterns across successive relaxations. Using these insights, we develop normalization techniques and supervised models to identify dominating states, both within the same problem and in new instances. Finally, we leverage these results to propose an initial design for a data-driven dynamic programming algorithm for both single and multi-resource problems, setting the stage for further refinements and extensions.
- Short bio: Saverio Basso is a researcher at the Dalle Molle Institute for Artificial Intelligence (IDSIA), part of the University of Applied Sciences and Arts of Southern Switzerland. He earned his PhD and MSc in Computer Science from the University of Milan. His doctoral research focused on data-driven techniques for generating Dantzig–Wolfe decompositions for mixed-integer problems, as well as on the design of parallel and distributed column generation algorithms. His primary research interests lie in combinatorial optimization problems, with a particular focus on exact and data-driven approaches to resource-constrained shortest path problems.
- Speaker: Niels Agatz (Erasmus University)
- Title: How Aesthetics influences Aversion of Routing Operations Algorithms (joint with Yingjie Fan and Daan Stam)
- Time: Thursday, Oct. 23, 2025, 13:15 - 14:45
- Location: Lecture Hall 8 (1-floor), Oskar-Morgenstern-Platz 1
- Abstract: Logistics service providers often have human planners interact with automated planning systems to plan their delivery routes. Anecdotal evidence suggests that planners may not use route plans that "don't look nice," even if they are optimal in terms of the number of vehicles, travel distances, or times. Although operations researchers have begun to incorporate aesthetic factors into vehicle routing models, there have been no empirical studies on the impact of aesthetic factors on route selection decisions. Here, we present the results of a systematic study examining the impact of visual appeal in vehicle routing plans based on behavioral experiments.
- Bio: Niels Agatz is a full professor at the Rotterdam School of Management, Erasmus University and the scientific director of TKI Dinalog, the Dutch national institute for collaborative logistics research. His research focuses on improving sustainable last-mile supply chains using analytics and technology-driven innovation. He earned his MSc from Eindhoven University of Technology and his PhD from Erasmus University Rotterdam.
- Speaker: Remy Spliet (Erasmus University Rotterdam)
- Title: A Partial Path Formulation of a Crew Scheduling Problem (joint with Bart van Rossum and Twan Dollevoet)
- Time: Friday, Oct. 24, 2025, 09:45 - 11:15
- Location: Seminar Room 5 (1-floor), Oskar-Morgenstern-Platz 1
- Abstract: We study a crew scheduling problem, in which tasks are allocated to crew to form duties in such a way that working time, duty length and break regulations are not violated. Traditional integer programming approaches to crew scheduling problems use formulations in which a variable is included for each of the exponentially many duties and column generation is used to solve the corresponding LP relaxation. Instead, we propose using partial duties, which simplifies the pricing problem that is iteratively solved by column generation algorithms. Although this comes at the expense of a weaker LP bound, we introduce valid inequalities that repair the value of the LP relaxation. We perform numerical experiments with a traditional branch-price-and-cut heuristic to assess the performance of the new approach.
- Bio: Remy Spliet is Associate Professor at the Econometric Institute of the Erasmus School of Economics, at Erasmus University Rotterdam, the Netherlands. His research is in the field of Operations Research with a focus on Transportation.
- Speaker: Dilay Aktas Dejaegere (KU Leuven)
- Title: Balancing Service Levels and Operational Efforts in Dynamic Waste Collection
- Time: Wednesday, Nov. 5, 2025, 16:30 - 17:30
- Location: OMDS Seminar Room (03.307, Oskar-Morgenstern-Platz 1)
- Abstract: Cities face growing challenges in keeping waste collection both efficient and responsive to citizen needs. Digital technologies such as sensor-equipped bins enable real-time information on fill levels and open the door to dynamic planning. In this talk, I present a dynamic optimization framework for smart waste collection that explicitly considers the trade-off between service levels and operational effort. I will also discuss how limited, strategically placed sensors can achieve performance close to full deployment, offering practical insights for cost-effective smart city solutions.
- Bio: Dilay Aktas Dejaegere is a postdoctoral researcher at KU Leuven Institute for Mobility, Belgium, where she also obtained her PhD. She has a background in operations research, studied design and optimization of demand-responsive public bus systems during her PhD, and now works on dynamic optimization frameworks for smart waste collection.
- Speaker: Sanne Wøhlk (Aarhus University)
- Title: Solution of a Truck-and-drone Routing Problem (joint work with Jahir Llagas and Marcel Turkensteen)
- Time: Wednesday, Nov. 26, 2025, 16:30 - 17:30
- Location: OMDS Seminar Room (03.307, Oskar-Morgenstern-Platz 1)
- Abstract: In urban logistics, the integration of drones with traditional delivery methods presents a promising avenue for improving efficiency and sustainability. This study addresses the optimization of last-mile deliveries by a single drone-truck system. We consider the case with one drone and one vehicle, which both start and terminate at the depot. The drone can make autonomous deliveries, can pick up packages independently, and can recharge by itself. The resulting problem, which we call a Single Drone-Truck Routing Problem, has the challenge of coordinating the drone and the truck while constructing routes. We propose an exact approach based on Logic-Based Benders Decomposition to solving the problem, which is both more efficient and more flexible than a straight-forward model for the problem.
- Speaker: Prof. Gudrun Kiesmüller (TUM Heilbronn)
- Title: Planning of Service Technicians for After-Sales Field-Service
- Time: Wednesday, Dec. 3, 2025, 15:00 - 16:30
- Location: HS 9 (1-floor, Oskar-Morgenstern-Platz 1)
- Abstract: To promote customer loyalty and generate revenue in after-sales service, many companies provide field repair services to their customers. Service technicians, who perform these repair jobs, may have different skills and experience, which can result in different probabilities of successfully completing a repair job. In case of a failed repair job, a second visit is required. We consider the planning problem of a service provider who must decide which repair tasks to allocate to a specific technician and in what sequence the technician should visit the customers. We assume that customer requests for service arrive randomly and that planning is performed on a daily basis. We formulate the problem as a stochastic sequential decision problem, where at the start of each working day, the information about unsatisfied customer requests is summarized, technicians are allocated to repair tasks, and visiting routes are planned. The objective is to minimize the expected total discounted costs, composed of technicians’ travel and customer waiting costs. We propose an innovative hybrid value function approximation method that efficiently handles large decision spaces. We conduct an extensive numerical study to investigate the performance of the method and to derive insights about good allocation decisions.
- Bio: Prof. Kiesmüller studied mathematics at Julius Maximilian University of Würzburg, where she also earned her doctorate. She then worked as a postdoctoral researcher and later as an assistant professor at Eindhoven University of Technology in the Netherlands. She was appointed full professor of supply chain management at Christian Albrecht University in Kiel and professor of operations management at Otto von Guericke University in Magdeburg. Since 2019, she has been full professor of operations management at TUM Campus Heilbronn. Prof. Kiesmüller's research focuses on supply chain management, inventory management, and, in particular, the management of spare parts inventories. Recent projects also include the planning of maintenance processes and the design of manufacturing systems. Her work includes the development of optimization approaches for decision support. Together with various co-authors, she has published in leading international journals such as IISE Transactions and Production and Operations Management.
