Automatisierte Umlaufplanung im öffentlichen Verkehr

Abstract

Schedule planning in local public transport is an NP-hard problem, also known as the
Multi-Depot Vehicle Scheduling Problem (MDVSP). Manual planning is inefficient and
expensive. The goal of this thesis is to develop and evaluate an automatic schedule planning system that incorporates historical delay data to generate more robust schedules. The
system aims to reduce operational costs and increase service reliability.
After reviewing the current literature, an Iterated Local Search (ILS) algorithm was
chosen to solve the MDVSP. The prototype implements a line-based solution representation and uses a Block-Moves neighborhood. Historical delay data are incorporated through
delay propagation risk analysis and strategic slack time allocation. The evaluated C# .NET
prototype was developed using real operational data from Austrian transport companies.
The results show a significant improvement over manual planning, reducing both fleet
size and operational costs. This demonstrates the potential of the developed ILS algorithm to make public transport more efficient and to increase service reliability through
an automated, data-driven planning system.

Date of Award	2025
Original language	German (Austria)
Supervisor	Herwig Mayr (Supervisor) & Iris Schlager (Supervisor)