Suchergebnisse

Verden har gennem de senere ar oplevet en stigende eftersprgsel efter transport og mobilitet. Med den voksende persontransport og den deraf forstrkede klimapavirkning bliver det kollektive transportsystem et glimrende alternativ til den individuelle transport, idet det kollektive transportsystem kan tilbyde omkostningseektiv, klima- og miljvenlig transport samt at dette er eektivt i kampen mod trngsel. I 2009 vedtog det danske Folketing transportaftalen: "En grn transportaftale", hvori det sttes som mal, at den kollektive transport skal udbygges og eektiviseres saledes, at den forventede vkst i persontrakarbejdet ville kunne lftes af det kollektive transportsystem. I Danmark har det kollektive transportsystem en markedsandel pa 13% af transporten mellem arbejde og hjem og der blev i 2010 gennemfrt mere end 290 millioner togrejser i det danske jernbanenet. Det er endvidere tidligere blevet estimeret, at der hver dag spildes 65.000 timer af passagerer, der venter pa forbindelser i den kollektive transport. For at imdekomme den forventede vkst og for at kunne tilbyde passagererne en attraktiv service til en fornuftig pris skal den kollektive transport vre sa eektiv som mulig. Forskningen har gennem de seneste mange ar demonstreret, at matematiske optimeringsmetoder kan gre det kollektive transportsystem mere eektivt og dermed mere konkurrencedygtigt i forhold til andre transportformer. Planlgning af kollektiv transport bestar af en rkke forskellige optimeringsproblemer pa savel det strategiske, det taktiske og det operationelle planlgningsniveau. Problemstillinger som for eksempel linjeplanlgning, kreplansplanlgning og kretjsplanlgning lses traditionelt uafhngigt og sekventielt. Der ndes i dag adskillige modeller og eektive metoder til lsning af disse problemstillinger. Lses disse problemer uafhngigt og sekventielt vil det saledes ofte resultere i lsninger, der vil opleves som vrende suboptimale for passagererne i forhold til serviceniveau. Lsninger der giver passagererne en bedre service kan opnas ved at integrere srligt relevante planlgningsproblemer. Grundet strrelsen af virkelige planlgningsproblemer, er det dog i midlertidigt, som oftest, ikke muligt at lse det integrerede problem uden srligt udviklede metoder, der endvidere krver meget store beregningsmssige ressourcer. Malet med denne afhandling er at studere integrationen af kreplansplanlgningen med andre planlgningsproblemer sasom linjeplanlgning, rutevalg og planlgning af kretjernes opgaver (omlbsplanlgning). Vi udvikler en rkke avancerede planlgningsvrktjer, der kan nde integrerede lsninger til de komplekse planlgningsproblemer man mder inden for den kollektive trak. Vi undersger anvendelsen af disse integrerede modeller til at lse virkelige casestudier, og vurderer kvaliteten af de fundne integrerede lsninger sammenlignet med lsninger baseret pa ikke-integrerede modeller. Afhandlingen beskftiger sig med to specikke problemstillinger, der ses i den kollektive bustrak og en problemstilling fra den kollektive jernbanetrak. I problemstillingen fra den kollektive bustrak integrerer vi kreplaner med bade kretjsplanlgning og passagerernes rutevalg. Forholdet mellem afgangs- og ankomsttider og passagerernes rutevalg unders ges under hensyntagen til omkostningerne ved at udfre kreplanerne. I problemstillingen inden for jernbanetrakken undersger vi delvis integration af kreplansplanlgning og linjeplanlgning. Konkret studeres, hvordan kreplaner kan forbedres med hensyn til denerede mal for den service, der ydes til passagererne. Disse forbedringer opnas ved at tillade ndringer i frekvenserne for togafgangene og gennem ndringer i standsningsm nsteret for de stationer der er inkluderet i ruten. Vi formulerer problemerne ved hjlp af matematiske modeller, og vi foreslar heuristiske lsningsmetoder til dem alle. Lsningsmetoderne testes i virkelige casestudier for Storkbenhavnsomradet for problemstillingen fra den kollektive bustrak og for en delmngde af intercity-netvrket i Holland for problemstillingen fra den kollektive jernbanetrak. Vi viser, at kreplanerne, der er opnaet med de integrerede modeller, resulterer i at servicen til passagerne ges samtidig med, at driftsomkostningerne kan bibeholdes pa et undret niveau eller holdes inden for et givent budget. ; Recent years have seen a general increase in the demand for transport and mobility. As the transportation numbers rise, and with the increased eorts in reducing emissions on a global scale, public transport becomes an excellent alternative to private transport due to being cost efficient, environmentally friendly, and for its abilities to reduce congestion. In 2009, the Danish Parliament passed the transport policy agreement entitled \A green transport agreement", which states that the public transport system should be expanded and made more ecient so that the expected growth in person transport should be able to be captured by the public transport system. In Denmark, public transport holds a market share of 13% of all commuting, and in 2010 more than 290 million passengers used the railway network. Furthermore, in Copenhagen it was estimated that 65,000 hours are lost every day by passengers waiting for connecting services in public transport. In order to serve increased demand and provide an attractive service for the passengers at an affordable cost, public transport should be both time and cost efficient. Research has through a number of years clearly shown that operations research techniques can help make public transport systems more efficient and thereby increase its competitiveness compared to other transport modes.Public transport planning comprises a range of planning problems at the strategic, the tactical as well as the operational planning levels. Problems like line planning, timetabling, or vehicle scheduling are traditionally solved independently and sequentially. Today, several models and efficient solution methods for solving these exist. However, solving these problems independently will often result in suboptimal solutions for passengers with respect to the service level offered. Solutions with increased passenger service could be obtained by integrating different planning problems. However, due to the size of real world instances, solving the integrated problem is, usually, not possible without dedicated solution methods, which, in turn, requires huge computational efforts. The goal of this thesis is to study the integration of timetabling with other planning problems such as line planning, passenger routing, and vehicle scheduling. We propose a number of advanced planning tools, which are able to find integrated solutions to the complex planning problems seen within public transport. We investigate the application of integrated models to real life case studies and we assess the quality of integrated solutionsin comparison to non-integrated solutions. The thesis studies two problems in the bus transport domain and one problem in the rail transport domain. In the bus transport domain, we integrate timetabling with both vehicle scheduling and passenger routing. The relations between departure and arrival times and how passengers travel in the network are investigated, while taking into account the costs associated with operating the timetables. In the train domain, we investigate partial integration of timetabling and line planning. Specically, how timetables can be improved in terms of passenger service by allowing changes in train frequencies and changes in stops at stations included in the train journey. We formulate the problems using mathematical models, and we propose heuristic solution methods for all of them. The solution methods are tested in real life case studies of the Greater Copenhagen area for the bus transport problems, and of a subset of the intercity network in The Netherlands for the rail transport problem. We show that the timetables obtained with integrated models provide an increased passenger service while being able to keep operating costs at the same level or within a pre-dened budget.