The OMDS/BALOR Talk Series invites you to the following seminar talks in June 2026:
- BALOR Scientific Talk: Towards industrial strength Roll-On Roll-off stowage planning. A bumpy ride.
- Speaker: Prof. Dr. Dario Pacino (Technical University of Denmark)
- Date/Time: Wednesday, June 24, 16:45 - 18:15
- Location: OMDS Seminar Room (03.307), Oskar-Morgenstern-Platz 1
- Abstract: Stowage planning is the art of deciding where cargo should be placed on a vessel. A Roll-On Roll-Off (RoRo) vessel is what most people would refer to as a ferry. You drive on at arrival (Roll-On) and drive off at departure (Roll-Off). Ferries are planned on the fly, and this is possible because most of the vehicles on a ferry have an easily foreseeable weight and size. A RoRo vessel for freight transport is similar, but cargo weights and size can vary drastically. Weight, in particular, can range from 10 to 80 tons. Placing an 80-ton paper-roll trailer wrongly, could drastically impact vessel stability, fuel consumption, and ultimately emissions. In this talk I will introduce you to the world of RoRo shipping while touching upon developments on stowage planning optimization. The talk will then deep-dive into recent and promising results on solution methods based on Reinforcement Learning. I will present open problems in RoRo shipping, and tease you a bit with the challenges of making a research roadmap in collaboration with the industry and funding agencies.
- BALOR Lab Talk: Heterogeneous Fleet Routing under Environmental Budget Constraints
- Speaker: Nicolas Kuttruff (TU Munich)
- Date/Time: Thursday, June 11, 15:00 - 16:30
- Location: SR 13 (2-floor), Oskar-Morgenstern-Platz 1
- Abstract: Urban freight distribution is increasingly shaped by city-level regulations, including low-emission zones, access restrictions, and policies promoting zero-emission delivery, forcing logistics providers to operate heterogeneous fleets that combine diesel and battery-electric vehicles. We study a large-scale heterogeneous vehicle routing problem with nonlinear operating costs and environmental budget constraints. Energy consumption, CO2 emissions, and operating costs depend jointly on vehicle type, payload, speed, and travel distance, implying that route costs cannot be represented as fixed arc attributes. Consequently, the cost and emissions associated with serving a customer depend on both the customer’s position within a route and the assigned vehicle type. We consider alternative regulatory formulations based on total emissions, emissions per kilometer, and diesel-distance ratios as flexible substitutes for fixed spatial restrictions such as electric-only zones. To solve the problem, we develop a decomposition-based matheuristic that clusters customers, generates candidate routes, evaluates them using load-dependent energy and emission models, and recombines routes through set partitioning. An adaptive multi-armed bandit learning mechanism dynamically selects decomposition and routing strategies during the search, balancing exploration and exploitation across different parameter configurations. The proposed approach outperforms the commercial solver Hexaly by up to 9.6% on the nonlinear objective for large-scale synthetic benchmarks within equal runtime limits. In a case study based on urban delivery operations in the Munich metropolitan area, the experiments further reveal clear trade-offs between cost and emissions. In particular, a flexible electric-kilometer budget reduces emissions by 16.6% relative to an electric-zone baseline without increasing total operating costs.
