Florian Fuchs


Last Name

Fuchs

First Name

Florian

ORCID

0000-0001-7563-0830

Organisational unit

09611 - Corman, Francesco / Corman, Francesco

Filters

2022 - 202521
Reset filters

Search Results

Publications 1 - 10 of 21
  • Handling conflicts in microscopic railway timetabling at scale
    Item type: Working Paper
    Fuchs, Florian; Dubach, Thomas; Lordieck, Jan; et al. (2025)
    SSRN
    This paper presents a hybrid optimization framework for microscopic railway timetabling that integrates Logic-Based Benders Decomposition with Conflict Discovery (CD-LBBD) and a Soft-Conflict Mixed-Integer Programming (SC-MIP) formulation. The CD-LBBD method separates high-level coordination from microscopic feasibility, employing an SMT-based subproblem to dynamically detect and resolve resource conflicts. The SC-MIP formulation, by contrast, introduces a continuous relaxation of disjunctive precedence constraints and iteratively minimizes conflict violations to obtain feasible schedules that are subsequently refined to optimality. Both approaches operate on a strengthened continuous-time event–activity formulation enhanced by bound propagation, structural clustering, traversal bounds, and resource aggregation into train-specific chunks.The proposed framework is evaluated on the complete DISPLIB~2025 benchmark set, comprising 112 real-world–based instances. Within a 10-minute competition limit, CD-LBBD and SC-MIP jointly solved 65 instances to optimality. The hybrid solver achieves 91 best-known solutions—approximately 81\% of the benchmark—and obtained the highest overall score in the DISPLIB~2025 competition. These results demonstrate that combining exact methods with complementary algorithmic structures provides robust and scalable performance for large-scale, microscopic railway timetabling.
  • DISPLIB competition
    Item type: Conference Poster
    Dubach, Thomas; Fuchs, Florian; Lordieck, Jan; et al. (2025)
    DISPLIB 2025 challenged participants to advance real-time railway scheduling. We present a hybrid approach: Logic-based Benders decomposition dynamically discovers and separates train conflicts, while an incremental MIP relaxes constraints and iteratively converges to optimal solutions. This dual strategy maintains tractability while providing optimality guarantees.
  • Optimizing periodic stability in railway timetables: A microscopic model for networks with a macroscopic comparison
    Item type: Journal Article
    Fuchs, Florian; Martin-Iradi , Bernardo; Corman , Francesco; et al. (2025)
    Journal of Rail Transport Planning & Management
    We present a novel microscopic model for railway timetabling designed to maximize periodic stability in mixed single- and multi-track networks. Unlike conventional approaches based on the Periodic Event Scheduling Problem (PESP), our model provides a detailed infrastructure representation with flexible routing and nuanced conflict resolution, enhancing adaptability to real-world constraints and facilitating practical implementation by operators. To ensure scalability, we integrate a Satisfiability Modulo Theories (SMT)-based approach, which efficiently narrows feasible cycle time bounds, enabling the model to handle large-scale networks. Validated on operational data from the Rhätische Bahn network—a Swiss railway with complex infrastructure—the microscopic model consistently yields lower minimal cycle times than its macroscopic counterpart. The comparative analysis also offers insights into the trade-offs between model detail, computational efficiency, and achievable cycle times across diverse operational scenarios. These findings underscore the importance of infrastructure abstraction and the careful consideration of operational and commercial interdependencies for optimal stability in complex railway networks.
  • What we learned so far with E-Bike City Research D-BAUG Lighthouse Project
    Item type: Presentation
    Elliot, Catherine; Axhausen, Kay W.; Marra, Alessio Daniele; et al. (2024)
  • Multi-scale responsive public transport for bi-modal demand
    Item type: Conference Poster
    Fuchs, Florian; Corman, Francesco (2022)
  • Herausforderungen und Chancen durch integrierte (Linien-)Planung
    Item type: Other Conference Item
    Fuchs, Florian (2022)
  • Routing passengers while timetabling based on promises from line planning
    Item type: Conference Paper
    Fuchs, Florian; Klasovitá, Viera; Corman, Francesco (2023)
    hEART 2023: 11th Symposium of the European Association for Research in Transportation
    Effective line planning and timetabling are critical for enhancing public transport efficiency and passenger satisfaction. We propose a Logic-Based Benders decomposition approach to optimise a timetable for a passenger railway system based on the promises made in earlier planning stages. Our approach ensures that the promised travel times and transfers are available and passenger routes are chosen according to the shortest available path. We test this approach on real-world data from the Rhätische Bahn railway system, demonstrating promising results. The proposed approach has shown to be valuable for optimising transfers, improving efficiency and passenger satisfaction, and reducing travel times. The method has limitations, including the inability to consider multiple connections per origin-destination pair, adaptation time at the origin station, and crowding. Further research can focus on improving and extending the model's performance to include these factors.
  • Joint optimal periodic timetabling and train routing
    Item type: Other Conference Item
    Fuchs, Florian; Corman, Francesco (2023)
    RailBelgrade 2023: Book of Abstracts
    Timetabling and selecting routes for trains are two inherently related topics, as one affects the possible solutions of the other. While considering the problems independently brings the advantage of comparably easier stand-alone problems, it might come at a high cost of wasted capacity and unsatisfactory quality. Thus, we consider both planning problems simultaneously. Besides, we aim to provide exact solutions and model infrastructure topologically accurate on a mesoscopic level. Besides accuracy, optimality is part of our scope. Thus, we model the problem as a mixed integer problem. However, as we combine two problems that are already challenging when considered independently, we provide several means to reduce the size of the instance and measures to strengthen the formulation. While these improvements positively affect computation time, they do not affect solution quality. Thereby, we conserve optimality. As a result, our approach provides simultaneously optimised timetables and train routes. We assess our proposition on real-life instances of varying size and network utilisation. The results underline the drastic effect of our strengthening propositions, as we report a speed-up of more than two orders of magnitude in small instances. For more extensive instances where no optimal solution can be found within the time limit, reducing and strengthening yields higher quality solutions and optimality gaps of less than half the value when compared against the non-reduced/strengthened formulation.
  • Addressing uncertain vehicle availability in line planning
    Item type: Other Conference Item
    Fuchs, Florian; Corman, Francesco (2023)
  • Efficient railway timetabling and vehicle circulation
    Item type: Other Conference Item
    Fuchs, Florian; Leutwiler, Florin; Corman, Francesco (2024)
    This paper examines the optimization of the Periodic Timetabling Problem (PESP) and Vehicle Circulation through a novel Satisfiability Modulo Theories (SMT) approach, applied to the Rhätische Bahn Network (RhB-Network). Like traditional Mixed Integer Programming (MIP), our methodology utilizes the Event Activity Network (EAN) to model railway timetable events and activities. However, our approach encodes the PESP and vehicle circulation constraints within an SMT framework, extending SAT with difference logic. This extension allows for high-resolution time calculations, addressing the scalability limitations of existing SAT methods. Furthermore, we introduce the concept of selectable activities, enhancing the model’s capability to incorporate vehicle circulation effectively during train scheduling and routing. Empirical analysis on the RhB Network reveals that although MIP and SMT solvers show comparable performance on smaller instances, the SMT solver significantly outperforms on larger scales, consistently finding optimal solutions. This study underscores the potential of SMT-based methods for complex scheduling problems, offering significant advancements in public transport optimization.
Publications 1 - 10 of 21