Quantitative Methods for Assessment of Railway Timetables

Abstract

The aim of this PhD thesis is to improve the attractiveness of future railway timetables. To achieve this goal,the qualitative term “timetable attractiveness” needs to be made quantifiable. To establish what timetableattractiveness is, the thesis gives an introduction to railway timetables in the form of a timetable definitionand an overview of commonly used timetable types and existing timetable classes. All major timetablestakeholders in a given railway sector must agree on the timetable aspects covered by the term “timetableattractiveness”. This research succeeded in creating such an agreement in the Danish railway sector,through a process that included individual stakeholder interviews followed by a joint timetabling criteriaworkshop. The agreement is a list of six prioritized timetable evaluation and optimization criteria.To make the evaluation criteria quantifiable, a set of key performance indicators (KPIs) was developed foreach criterion. A total of 13 KPIs are presented. Their practical applicability has been successfully tested onexamples of real-life Danish timetables. The thesis recommends a revised timetabling process at railwayinfrastructure manager Rail Net Denmark (in Danish: Banedanmark) that will take the recommended KPIsinto consideration. This new time-tabling process will create the necessary foundation for improving futurerailway timetable attractiveness in Denmark.The aim and the structure of this PhD thesis are described in Chapter 1. The thesis shows how railwaytraffic and the railway timetable have been interlocked almost from the opening of the first public railway linein 1825. A timetable increases both the traffic safety and the attractiveness of railways. Timetables plantraffic, avoid train collisions, and announce train services to potential customers.Chapter 2 gives an introduction to railway timetables. The thesis identifies a need for an improved Europeandefinition of a railway timetable. A two-part definition is introduced, with one part covering a liberalizedrailway sector and the other covering a state-owned or completely privately owned railway monopoly. Theformer is an agreement between an infrastructure manager and one or more train operating companies. Thelatter is an internal company agreement.This is followed by the presentation of some basic facts that apply to all railway timetables. A railwaytimetable has a time period of validity. In the European Union this has been harmonized to one year and thetimetable changes take place on the Sunday following the second Saturday in December.The thesis distinguishes between timetables created as part of long or short-term planning. A short-termtimetable tries to make the best of the available resources, such as railway infrastructure and rolling stockfleets, to accommodate as many railway customer preferences as possible. When preparing a long-termtimetable, the starting point is to create an ideal timetable that fulfils the customer preferences on theassumption that the necessary infrastructure and rolling stock can be made available.Moreover, the thesis identifies eight basic line structures for train services found in a railway timetable. Thesebasic train service line structures are:1. Point-to-point line (high speed trains)2. Circle line (suburban trains or metro)3. Centralized nodes (e.g. Paris or London)4. Corridor / tree structure (e.g. the Danish railway network)5. Star shape (e.g. the metro in Rome)6. Universal star shape (e.g. the metro in Athens)V7. Railway network with a core route (suburban trains in Copenhagen and Munich)8. Meshed railway network (German InterCity-Express trains).Chapter 3 gives an overview of the six most commonly used timetable types in the railway sector. Thesetimetable types are:1. The public timetable (available to everybody, on paper and/or digitally)2. The working timetable (primarily used by train staff)3. The graphical timetable or train graph (used by timetable planners and traffic dispatchers)4. The track occupation diagram (used by timetable planners and traffic dispatchers)5. Rolling stock roster plan (used by employees of the train operating companies)6. Train staff roster plan (used by employees of the train operating companies).The thesis identifies seven existing basic railway timetable classes. At the beginning of Chapter 4, anextended timetable categorization model is presented. The categorization of timetables into classes is basedon the level of structure in a given timetable. A set of basic structural characteristics determine the level ofstructure. These structural characteristics are:x Timetable periodicity/the timetable is systematic (repeating traffic patterns)x Timetable symmetry (same stopping pattern and travel times for both driving directions of a trainservice)x Train meetings at selected station hubs (optimal transfer options)x High frequency train services (train services run at least every 10 minutes)The seven basic timetable classes identified by this thesis are:x The periodic /systematic timetable- The symmetric periodic /systematic timetable- The integrated fixed interval timetable (IFIT)x The high frequency timetablex The non-periodic/non-systematic timetable- The symmetric non-periodic/non-systematic timetable- The integrated non-periodic/non-systematic timetableBased on earlier British and Swiss approaches to measuring the level of structure in a railway timetable, thethesis proposes improvements to these existing methods and introduces two new timetable structureindexes based on the newly developed concept of timetable patterns. The two indexes are:x Systematic timetable index – using the most used timetable pattern timewisex Systematic timetable index – using the longest continuous timetable pattern timewiseFinally, the seven timetable classes are described and compared in detail with each other. This gives anoverview of strengths and weaknesses based on selected, but generally accepted, timetable evaluationcriteria.Identifying the basic railway timetable classes in a timetable that covers an entire network is difficult andlabour-intensive work. Chapter 5 presents a series of examples of timetable analyses to illustrate theVI Summarycomplexity of the task. It starts with the example of one railway line section served by one train servicerunning according to one timetable class; goes on to the more complicated example of one railway linesection served by several train services with different timetable classes, and ends with the most complicatedexample of an entire railway network served by several timetable classes.The thesis recommends weighting the timetable classes identified with timetable statistical factors such as:the number of train runs, passenger numbers, freight tons, train-kilometres, and passenger or freight tonkilometres.Chapter 6 starts with a brief historical overview of the liberalization process in the European railway sectorand of the liberalization of the Danish railway sector in particular. This is followed by a presentation of theoverall Danish railway timetabling process. The process is one of collaboration between the infrastructuremanager Rail Net Denmark (in Danish: Banedanmark), the train operating companies, and the DanishTransport Authority (in Danish: Trafikstyrelsen), which is the buyer of public railway service traffic.Next comes a detailed description of the timetabling processes at the following railway timetablestakeholders:x The Danish Transport Authority – buyer of public railway service trafficx The train operating company DSB – the largest passenger train operatorx The state-owned infrastructure manager Rail Net Denmark – prepares the annual timetableThere is no formal timetabling process at the Danish Transport Authority. It changes from project to project.Since Rail Net Denmark is a member of the professional body of European infrastructure managers, Rail NetEurope, the basic timetable process steps and their deadlines are already given for the annual nationaltimetable. Surprisingly, the basic Rail Net Europe timetabling process has no built-in formal learning loop.Both DSB and Rail Net Denmark have informal learning loops in their existing timetabling processes, in theform of experience based input from employees at the beginning (Rail Net Denmark) and/or evaluation of theproposed timetable before the final approval (DSB and Rail Net Denmark).The research for this thesis initiated a process to reach a consensus on timetable attractiveness in the formof timetable evaluation and optimization criteria in the Danish railway sector for the first time ever. Chapter 7describes the process in detail. First, the most important railway timetable stakeholders were identified. Theyare:x DSB – the biggest Danish passenger train operating companyx Arriva Danmark –the winner of the first public passenger traffic tender in Denmarkx DB Schenker Rail Scandinavia – the biggest freight train operating companyx The Danish Transport Authority – the buyer of public railway service trafficx Rail Net Denmark – the state-owned infrastructure managerThe process started with an individual interview with each stakeholder to establish a prioritized list of the fivemost important timetable evaluation criteria in the opinion of the interviewee. This was followed by a jointtimetabling criteria workshop based on the five lists of prioritized criteria from the interviews. Arriva Danmarkwas not able to participate in the workshop. The participants of the workshop reduced the number oftimetable evaluation criteria to six on their own initiative through discussion and dialogue. To achieve anVIIindividual ranking of the six criteria, each stakeholder was given three votes and was asked to give threedifferent criteria one vote each. The result of the voting was three layers of priority with two criteria in eachlayer:x High priority (3 votes): Systematic timetables andCapacity consumption on railway line sectionsx Medium priority (2 votes): Robustness of the timetable andSocietal acceptance of the timetablex Low priority (1 vote): Attractive transfer options andTravel timesThe workshop was unable to achieve consensus on an individual ranking of the six selected timetableevaluation and optimization criteria, so this is the result of the process initiated and conducted in theresearch for the thesis. It is the first version of a common list of prioritized railway timetable evaluation andoptimization criteria in Denmark.Since the Danish railway sector is highly affected by the ever-changing national political climate, this list isnot necessarily very stable. The thesis therefore recommends that a similar (and perhaps improved) processshould be carried out every two to five years to ensure an up-to-date common understanding of timetableattractiveness in the Danish railway sector.A lack of focus on customer preferences was also identified through the results of the stakeholder interviewsand the workshop. None of the parties set aside enough resources to perform large analyses within thisimportant subject.Chapter 8 analyses each criterion from the common Danish list of prioritized railway timetable evaluationand optimization criteria with regard to the most important influencing factors. This includes the societalaspect in the form of political decision makers and railway customers. Technical aspects are covered in theform of railway train operating companies and infrastructure managers. The most important influencingfactors are shown to be “Political requirements”, “Customer requirements”, “Train operating companyrequirements” and “Infrastructure characteristics”.This thesis recommends eight new steps of analysis in a future timetabling process to ensure an improvedrisk and attractiveness evaluation of a timetable. It also presents a proposal for a revised timetabling processat railway infrastructure manager Rail Net Denmark. The basic working steps remain the same, since theyare given by European Union legislation and Rail Net Europe guidelines. The important changes are thattimetable planners will work with several timetable variants simultaneously and that a real iterative capacityallocation process with the train operating companies will take place. This will require a much more intelligentand efficient timetable planning system than is available today.A set of 13 key performance indicators for the Danish railway system is presented in Chapter 9. Seven ofthese are newly developed. Each key performance indicator is connected to one of the six timetableevaluation criteria. The key performance indicators are:1. Systematic timetable index (Systematic timetable)- Based on the total time of the most used timetable pattern2. UIC 406 methodology (Capacity consumption on railway line sections)VIII Summary- Compressed timetables based on the blocking time theory3. Degree of deviation from timetable planning rules (Robustness of the timetable)- Focus on agreed upon running times and timetable supplements4. Conflict Risk Index (Robustness of the timetable)- The number of potential train path conflicts at a station and their estimated risk level5. Timetable train path fix points (Robustness of the timetable)- Geographical and temporal distribution of potential train path conflicts6. Proportion of train paths with shared rolling stock (Robustness of the timetable)- Number of train paths with shared rolling stock compared to the total number7. Proportion of train paths with shared train staff (Robustness of the timetable)- Number of train paths with shared train staff compared to the total number8. Proportion of buffer time in turnaround time and hand-over time for rolling stock (Robustness ofthe timetable)- Level of time supplements at terminus stations for rolling stock until next departure9. Proportion of buffer time in turnaround time and hand-over time for train staff (Robustness of thetimetable)- Level of time supplements at terminus station for train staff until next departure10. Independent organization carrying out customer satisfaction surveys (Societal acceptance ofthe timetable)- Inspired by the British organization “Passenger Focus”11. Proportion of timetable transfer time prolongation (Attractive transfer options)- Timetabled extra transfer time compared to physical minimum possible transfer time12. Proportion of optimal transfer conditions (Attractive transfer options)- Number of transfers planned to take place on the same platform out of the total number13. Proportion of timetable travel time prolongation (Travel time)- Timetabled extra travel time compared to travel time for theoretical non-stop trainThese key performance indicators have proven themselves in practical applications on examples of real-lifeDanish timetables. All calculations were done manually, but they could be automated and integrated into future versions of timetabling software packages.