Fully realistic multi-criteria timetable information systems [Elektronische Ressource] / von Mathias Schnee

Dr.F14.07.1979ullyhenRealisticDr.Multi-CriteriahTimetablehscInformationhSystemsKVmannomagF200acD-17hDernbt:ereicWht:InformatikM?ller-HannderderT08.09.2009ecmhnisc29.10.2009heHoUnivulkersit?tamDarmstadtinzurbacErlangungReferendesProf.akKarstenademeiheiscoreferenhenProf.GradesMatthiaseineseDr.Datumrer.Einreicnat.ung:genehmigteTDissertationderv?ndliconPr?fung:HerrnDarmstadt,Dipl.-Inf.9MathiascSchhneeennzier:geborenAbstractMillions of people use public transportation and consult electronic timetable informa-tion systems. A customer selects from the connections offered by the system accordingto personal preferences. The chosen connection is typically a compromise based on theimportance of several criteria, including departure and arrival time, travel time, comfortand ticket cost. Consequently, multi-criteria optimization should be used to deliver “at-tractive” alternatives. We developed the concept of advanced Pareto optimality as anevolution of the classical Pareto optimality approach. It delivers more alternatives andremovesunattractivesolutionsfromtheresultstosuitthenotionofattractiveconnectionsfor all potential customers.Realistic modeling of the search for attractive connections leads to shortest-path al-gorithms. Fast search algorithms are needed to answer customer requests in only a fewmilliseconds since the schedules are modeled as large graphs (several hundred thousandedges and nodes).