Dynamic Event-Activity Networks in Public Transportation

Abstract

Real-time timetable information and delay management in public transportation systems are two challenging applications which can be modeled as optimization problems on dynamically changing, large and complex graphs, so-called event-activity networks.We describe both applications in detail, review the state-of-the-art and explain the requirements for systems solving these problems in a productive environment. Focussing on recent research on decision support for train dispatchers, we sketch the system architecture for the software prototype PANDA.