ABSTRACT. The Swedish railway sector has been in the forefront in Europe when it comes to regulatory reforms, organisational changes and market opening since the late 1980s. This plenary presentation gives an overview of the changes, the various steps in the reform process, and the current organisation. The experience and effects of the reforms (both positive and negative) are presented in terms of a number of developments: infrastructure investments, national and regional demand for rail services, market entry, tenders, prices, punctuality, safety, speed and capacity etc. An overall assessment summarizes the pros and cons and some current issues that remain to be handled.

ABSTRACT. With high frequency and unavoidable disruptions, metro systems are nowadays undertaking great emphasis on disruption management. This paper proposes a mixed integer programming model for train rescheduling in high-frequency metro systems during partial blockages. Several train rescheduling strategies are formulated into the model that considers station capacity and rolling stock circulation. The model is applied to a busy line of the Shanghai metro network. The computation time meets the real-time application requirement. The case study presents different influences of various disruption scenarios and emergency train constraints on the optimal solution.

ABSTRACT. Providing punctual, reliable and enough services to customers is one main goal of railway network operators. By automation of train scheduling, it is possible to schedule and route trains on the network closer to its maximal capacity, which is of great value for network operators. In this survey we state the general formulation of the railway scheduling problem and show the principle of decomposition as a way to tackle it. The literature shows many different decomposition approaches. With a survey we aim to summarize existing research and state possible new directions for future research.

ABSTRACT. This paper aims at improving the computational efficiency of an integrated optimization problem (PC) for large-scale cases. Two decomposition methods are considered, namely a geography-based decomposition and a train-based decomposition. We propose an integer linear optimization approach to implement the geography-based decomposition of the integrated optimization problem. With the decomposition, a number of sub-problems (that correspond to partitions of a railway network or trains) with couplings are obtained. We further introduce three distributed optimization approaches to deal with the couplings among the sub-problems. Experiments are conducted to examine the performance of the three proposed distributed optimization approaches, in terms of feasibility, computational efficiency, solution quality, and (estimated) optimality.

ABSTRACT. Rail transportation plays an important role in the rapidly changing multimodal transportation market for its stability and reliability, which is vital for train operating companies to maintain its competitiveness. Serving different kinds of passengers in emergency situation is a reflection of reliability of rail transportation. While unavoidable stochastic perturbations (e.g. bad weather) disrupt air transportation causing interruptions, rail rescheduling demand is produced deriving from the interruption of air transport, which defined as Unpredicted Large Passenger Flow (simply for ULPF) in this paper. The problem encountered by dispatchers is to reschedule trains to serve ULPF. We address the optimization problem of rescheduling trains to serve ULPF causing by interruptions of air transportation. In this paper, we focus on the rescheduling of High-speed Railway to evacuate ULPF, since the characteristic of air transport passengers, which are willing to pay high cost for short time, is different from other modes. Three dispatching strategies, which are organizing the seats remained, inserting new trains and the combination of the above two with transfer (e.g. passengers arrive at the destination through transfer between existing train and inserting new train), are used to reschedule ULPF in this paper. Moreover, an Integer Linear Programming (ILP) model is constructed for rescheduling trains to serve ULPF. The proposed model is solved by a standard ILP solver.

ABSTRACT. This is (due to sickness) not the final paper. The paper (including the abstract) will—according to agreement—be finalised as soon as possible.

ABSTRACT. Many HSR lines in China are heavily utilized, and it is desired to declare the bottlenecks in the railway network. UIC Code 406 is an easy and effective way of calculating the capacity consumption (UIC (2004)). Based on the UIC 406 capacity method, the capacity consumption of railway infrastructure can be measured by compressing the timetable graphs. The paper analyses the characteristics of Chinese railway operation and the features of the compress timetable method proposed by UIC 406.Regarding the UIC 406 capacity leaflet as a framework, the paper builds a calculation and assessment framework for Chinese HSR capacity, presents a capacity calculation model based on the compress timetable method and determine the capacity bottleneck in the rail. In this paper, the UIC 406 based capacity calculation model is applied to evaluate the capacity consumption of Beijing-Guangzhou HSR in China. The train operation plan should be rescheduled in the bottleneck section for a better capacity utilization, taking the passenger demand and the cooperation with the whole train operation into account. Reference to the compress timetable method in UIC 406, a train rescheduled model is established, with an objective to minimize the occupancy time of trains in the bottleneck section, considering headway constraints, dwelling time constraints, interstation travel time constraints and so on. Then a precise algorithm called Branch and Bound Algorithm is used to solve to model.

ABSTRACT. Railway capacity planning aims to determine the amount of traffic that can be operated on a given infrastructure. The timetable compression method described in UIC Code 406 has become one of the standard tools in this area. Motivated by the Swedish Transportation Administration's timetable independent adaptation of the methodology and its need for extension we explore how the compression method can be applied to evaluate the capacity of the underlying infrastructure for strategic planning rather than the occupation ratio of a specific timetable. By performing ensemble averaging of scheduled train sequences we abstract from a single timetable concept and perform a distributional analysis of timetable utilization. The methodology is applied in capacity assessment of railway stations and line segments. To mitigate decomposition-induced underestimation of network effects the compression area is extended and approaches to include interdependencies between stations and lines are investigated. The methodology is tested in a case study based on data from the Swedish Southern Mainline rail corridor.

ABSTRACT. Railway capacity is subject to various conditions, such as the minimum headways, the number of platforms and the number of rolling stocks and crews. This paper extends the railway capacity estimation problem by considering the electrical multiple unit (EMU) circulation. A MIP model based on a hybrid time-space-state network is applied to optimize the capacity utilization with the objective function of maximizing the number of desired trains. To overcoming the computation difficulties on a very large scale problem, a Lagrangian relaxation-based approach is proposed. This decomposes the model into timetabling sub-problem and EMU circulation sub-problem by dualizing the consistency constraints. We use the data of Beijing to Tianjin intercity railway to show the effectiveness of the approach. The experimental result shows that the total throughput of the railway system depends on either the infrastructure capacity or the EMU circulation according to different fleet sizes. A multi-objective Pareto analysis is conducted for analyzing the trade-off between different type of trains. The benefit of the decomposition solution approach is displayed by the performance comparison with the centralized and sequential method.

ABSTRACT. Railway timetabling is a prominent research area in railway research. The time table is usually shown as a time-space diagram. However, even algorithms that try to adapt/add to an existing timetable rely on mixed integer programming, on graph theory etc., but do not use the geometric representation of the time table. In this paper, we consider the problem of determining residual train paths (e.g., for freight trains) in an existing time table close to operation. We aim to restrict possible disturbance on existing (passenger) traffic, and, hence, insert train paths of a specified minimum temporal distance to other trains. We show how we can use algorithms for thick paths in polygonal domains to compute the maximum number of trains with a specified robustness to insert.

ABSTRACT. In railway lines in which both rapid trains and regular trains have to be operated and the frequency is very high, rapid trains inevitably have too much running time supplement. This means drivers of rapid trains have too much freedom in driving. Thus, if drivers of rapid trains do not run the train “properly,” delays occur for the train or for the succeeding train. Hence, in order to realize high punctuality, rapid trains must run so that they do not stop nor do not give an influence to the succeeding train, which is almost impossible if no information is given. In this paper, we propose an idea to specify not only arrival and departure times but running speed of rapid trains at critical sections in a timetable to make the train operation more punctual. In order to realize this idea, we settled two issues. One is to establish a method to decide the location and the running speed there in the timetabling process. The other is to implement a system which shows the speed information to drivers in an economical way. We have applied this approach when we revised our timetable in March 2018 and from the analysis of the historical train traffic data for several months, we have confirmed that the delays were greatly reduced and our approach was very successful.

ABSTRACT. This paper considers the problem of minimizing travel times and maximizing travel time reliability, which are important socio-economic properties of a railway transport service, for a given set of departures on a double-track line. In this paper travel time reliability is measured as the average delay, and a delay prediction model for MILP timetable optimization is presented. The average delay prediction model takes into consideration time supplements, buffer times and propagation of delays in the railway network and is not restricted to a fixed order of the trains. Validation of the average delay prediction model, and an evaluation of the approach with combined simulation-optimization for improving railway timetables, are conducted by a simulation study on a part of the Swedish Southern Main Line. Results from the simulation study show that the average delays are reduced by up to approximately 40% and that the punctuality is improved by up to approximately 8%.

ABSTRACT. The expansion of the scale of high-speed railway networks and the growth of passenger demand imply a high frequency of high-speed trains in China, i.e. higher railway capacity utilization. Based on given infrastructures and train line plans, there are some timetabling strategies which affect the capacity utilization, e.g. changing train departure sequence at origin stations, overtakings between trains, and adding new train stop at stations. Nowadays, managers of high-speed railway in China are eager to find out that what kind of impact these strategies have on the capacity utilization. In this study, new variables of train stops and constraints of overtakings are proposed with an extended cyclic timetabling model based on the periodic event scheduling problem (PESP). Minimum cycle time, train travel time and the total number of train stops are calculated as objectives to measure the differences between the strategies. The effectiveness of the three timetabling strategies are compared and presented by a series of experiments based on one real-world rail line in China. According to our results, with flexible train departure sequence at the origin stations and train overtakings, the possibility of acquiring good capacity utilization can be higher, but too many overtakings will have negative effect on the quality of timetable. The effectiveness of adding new stops on the capacity utilization depends on the ways of adding stops, i.e. which train is allowed to be added new stops and which stations can be selected to stop at.

ABSTRACT. Stockholm’s Metro is about to be expanded. Nearly twenty kilometres of new track and eleven new stations on four sections. The construction is planned to start in 2018/2019.

The purpose of this study was to evaluate the metro system performance for two different timetable scenarios. At the same time it should be evaluated if there was any significant impact on the simulation results whether Sofia station, where trains from different branch lines merges, is designed with two or three platform tracks.

Simulations were performed in RailSys with distributions prepared from log data regarding run times, dwell times and deviation from scheduled departure times. In order to fully model the signaling system behavior and setup regarding the tracks vertical profile around Sofia station a model was developed for this.

The results concluded that there could not be observed any significant effect on the expected delays during normal operations whether Sofia station consisted of two or three platform tracks.

ABSTRACT. The train service scheme of an urban rail transit line specifies information such as the total number of train services operated in the line, and the associated turn-back stations, train size and frequency of each service. A reasonable train service scheme can provide satisfactory services for passengers and reduce the operational cost for operators. This paper focuses on the optimal train service design problem in an urban transit line, where both the short-turn services and the train size of each service are considered. A service network based on a given pool of candidate train services with provided turn-back stations is constructed. The optimal strategy is used to assign passenger flows on the service network so as to describe the transfer process of passengers between different train services. Considering many operational and capacity constraints, a mixed integer nonlinear programming model minimizing the sum of the operators’ cost and passengers’ waiting time cost is developed to identify train services from the service pool and determine the train size and frequency of each chosen service. The nonlinear model is transformed into a linear one, and two simplification methods named service network simplification and OD pair aggregation are proposed to improve further the computational efficiency of the model. Finally, realistic instances from Chongqing Rapid Rail Transit Line 26 in China are used to test the proposed approaches. The results show that our approach can effectively reduce the operators’ cost and the passengers’ waiting time cost compared with the empirical method frequently used in practice.

ABSTRACT. With the rapid development of the Chinese high-speed railway (HSR) network, more and more railway lines are becoming oversaturated, leading to inefficient operations and reducing the service quality. To improve the network’s performance, this paper proposes a profit-oriented line planning model for collaboratively optimizing the operational costs and passenger travel times. Due to the complexity of the problem, an iterative approach is designed to solve the problem efficiently. Two case studies are implemented to verify the performance of the approach. The results of the small example show that the best found line plan can save up to 6% of the travel time compared to the initial solution and improve the profit with 2%. The proposed iterative approach also performs well in searching for high-quality solutions on the large railway network.

ABSTRACT. Multi modal journeys often involve two trips: one outgoing and one return trip, as in many cases, the traveller would like to return to his starting point. If a car or bike was used in combination with public transportation (i.e. park \& ride), this introduces a dependency between outward and return trip: both must include the same parking place. Optimizing both trips independently may yield suboptimal results. We consider the multi modal two-way roundtrip problem and propose several algorithms. All proposed algorithms compute journeys that are optimal regarding multiple criteria. Our study with realistic scenarios based on real data shows promising results.

ABSTRACT. Primary delays (PDs) are the driving force of delay propagation. Hence, accurate predictions of the number of affected trains (NATs) and the total time of affected trains (TTATs) due to PDs can provide a theoretical background for the dispatch of trains in real time. Train operation data were obtained from Wuhan-Guangzhou High-Speed Railway (HSR) station from 2015 to 2016, and the NAT and TTAT influence factors were determined after analyzing the PD propagation mechanism. The NAT predictive model was established using eXtreme Gradient Boosting (XGBOOST) algorithm which was more efficient than other machine learning methods after comparison. Furthermore, the TTAT predictive model was established based on the NAT model using the support vector regression (SVR) algorithm. The results indicate that the XGBOOST algorithm has good performance on the NAT predictive model, whereas SVR is the best method for the TTAT model using Lessthan5 variable, which is the ratio of the difference between the sample size of actual and the predicted values in less than 5 min and the total sample size. In addition, 2018 data were used to evaluate the application of NAT and TTAT models over time. The results indicate that NAT and TTAT models have a good application over time.

ABSTRACT. In our previous work a Markov chain model for the stochastic process of train delays at stations was presented (Şahin, 2017). The model proposed a method for developing the stochastic one-step transition matrices for the running time delays (for recovery) and the conflict resolution delays (for deterioration) using the real-life train operation data. It is possible to estimate the delay distributions at stations under the effects of running time supplements and buffer times. Hence, the model can be used to evaluate the effectiveness of the time supplements embedded into the train schedules, in addition to some other performance measures. In this current investigation, the use of the stochastic matrices is expanded to predict train delay states at the stations ahead along the train path. We investigate the effects of the recovery and deterioration transition matrices developed in combination with the current delay state of a particular train in order to predict its expected time of arrival at subsequent stations. The recovery and deterioration matrices are used in the prediction repeatedly (as many times as needed) whenever a new prediction is made. The former matrix is considered in the prediction when a train run occurs and the latter one is considered when a train is expected to interfere with another train. In addition to the delay and trajectory prediction, the model can also be used to trace delay propagation and to measure schedule robustness. The goodness of the model predictions is assessed against the real-life data for train movements.

ABSTRACT. The uncertainty of estimating the railway passenger flow in advance may disrupt the passenger operation and management (e.g., passenger evacuation planning, seat allocation, and train timetable programming). In order to proactively improve the service quality and efficiency of the railway system, the short-term passenger flow prediction technique is vital in the field of operation and management system. Utilizing the deep learning library-keras, the study develops a Long short-term memory neural network (LSTM NN) to predict the short-term high-speed rail (HSR) passenger flow. Processing the raw data, we first construct the passenger flow sequences as the input (output) variables. Then the gird search and cross validation techniques are applied to optimize the LSTM NN parameters. At last we utilize the data provided by Shanghai railway administration of China as the case study. Through a comparison with other representative methods, including Auto-Regressive Integrated Moving Average (ARIMA), Back Propagation Neural Network (BPNN), and Support Vector Machine Regression (SVR), results suggest that the proposed LSTM NN can generate great potentials for accurate passenger flow predictions.

ABSTRACT. In passenger railway stations, train units preparation is crucial for service quality. This preparation includes maintenance check, cleaning, coupling and uncoupling. Such operations require parking train units on shunting yards located close to platforms. Therefore trains have to be moved between platform and shunting tracks. Taking over train units between their arrival and their departure in a station constitutes shunting. The Generalized Train Unit Shunting problem (G-TUSP) is the problem of shunting operations planning. The problem is to assign arriving train units to departing train units, shunting tracks and paths, to schedule shunting movements and to assign crews to maintenance operations. The aim of the paper is to provide an algorithmic approach for the G-TUSP. The contribution presents an integrated problem with a mixed-integer linear programming (MILP) formulation. The formulation is based on a microscopic model of the infrastructure and formal train units in order to consider coupling and uncoupling. The model is solved exactly using the commercial solver CPLEX. It is tested on instances based on Metz-Ville station in France. The results are promising and show the suitability of the model.

ABSTRACT. Railway operators must schedule resources such as rolling stock and crew in order to operate trains as defined by a timetable. This paper considers scheduling of rolling stock, which is usually done by creating a roster. A roster is a series of trains to be performed by the particular rolling stock. The number of train-sets required to operate a given group of trains is essentially determined by the roster and generation of an efficient roster is essential. Important considerations of the roster generation include maintenance such as pre-departure inspection. On some lines in Japan, splitting and combining are often used to adjust transportation capacity flexibly. Under this type of operation, splitting and combining become necessary. These shunting operations require time and manpower, so it is necessary to reduce the amount of splitting and combining. This paper presents a mixed integer linear programming model so that the amount of splitting and combining is reduced together with the roster length and the distance of empty runs. Results of computational studies will be presented based on real instances of several lines in Japan, indicating the computational effectiveness of the methodology and with respect to the reasonableness of the resultant rosters.

ABSTRACT. In railway scheduling, a nominal traffic schedule is established well in advance for the main resources: train-paths, rolling stock and crew. However, it has to be adapted each time a change in the input data occurs. In this paper, we focus on the costs in the adaptation phase. We introduce the concept of adaptive nominal solution which minimizes adaptation costs with respect to a given set of potential changes. We illustrate this framework with the rolling stock scheduling problem with scenarios corresponding to increasing demand in terms of rolling stock units. We define adaptation costs for a rolling stock schedule and propose two MILPs. The first one adapts, at minimal cost, an existing rolling stock schedule with respect to a given scenario. The second MILP considers a set of given scenarios and computes an adaptive nominal rolling stock schedule together with an adapted solution to each scenario, again while minimizing adaptation costs. We illustrate our models with computational experiments on realistic SNCF instances.

ABSTRACT. With the booming travel demands in the urban cities, which can’t be satisfied due to the limited transportation capacity in urban rail transit, passenger congestion problem become increasingly serious, causing the potential accident risks on platforms. To further efficiently improve the conditions, this paper proposes an effective collaborative optimization method for the accurate passenger flow control strategies on an oversaturated urban rail transit line by simultaneously adjusting the number of inbound passengers entering multiple stations on the line. Through considering the space-time dynamic characteristics of passenger flow, a multi-objective mixed-integer linear programming model is formulated to firstly minimize the number of passengers who are limited to enter stations, secondly minimize the total passenger waiting time on platforms at all of involved stations where the optimal passenger flow control is imposed to avoid congestion on platforms within the transportation capacities, and thirdly maximize the passenger person-kilometres. Due to a small scale of the model, it can be solved by CPLEX solver efficiently. Moreover, because the passenger flow demand is time-variant, it’s very necessary for an accurate and easy-to-implement passenger flow control strategy to determinate the control time intervals. Hence, in order to get an optimal determination of the control time intervals, Fisher optimal division method is firstly applied to after modelling. Finally, two sets of numerical experiments, including a small-scale case and a real-world instance with operation data of Chengdu metro system, are implemented to demonstrate the performance and effectiveness of the proposed approach.

ABSTRACT. Optimizing the pedestrian facilities plan in transfer stations is the problem of adjusting the facilities on the layout of pedestrian flow route and the number of machines in service to service to meet the level of services requirements. In the practice, the operation of pedestrian facilities plan is always associated with the staff assignment. Hence, we develop a machine learning based integrated pedestrian facilities planning and staff assignment optimization model in transfer stations to schedule the pedestrian facilities plan and the staff assignment together. It aims to minimize the staff assignment cost and the deviation of working time of each employee of the station. The minimizing of the deviation gains the fairness of the assignment plan. The facilities plan is enforced by the level-of-services requirement in three performance indicators including transfer capacity, transfer average time and level-of-service. The performance indicators of facilities plans are evaluated by a simulation-based machine learning method. Based on simulation results, the random forest method fits a quantitative relationship among performance indicators of the facilities plans with operation scenario attributes and facilities plan attributes. The experiments on the case study of Xipu station show the integrated model can return pedestrian facilities plans which meet the level of service requirements and assign employees fairly of each period in a day and minimize the labor cost. The solutions of pedestrian facilities plan and staff assignment plan for possible operation scenarios in future are also suggested to station manager by our integrated method.

ABSTRACT. We built a test tool to estimate and visualize the congestion of the station. We utilized only existing data, we didn't install new sensors. Since it is a test tool this time, we do not actually expect congestion in real time. We gathered past data and verified whether it was predictable or not.

In this tool, we estimated by loading weight data, delay data, and ticket gate data. From these data, we calculated congestion at each place of the station. Since the data that can be acquired in real time is only the loading weight data and the delay data, regarding ticket gate data, we estimated real time congestion based on past data.

Since this research is in its early stages, we selected two stations that are easy to estimate. At these stations, we divided the areas by platforms, stairs, and concourses and made an estimate.

Regarding the display method of the estimation result, we displayed three-level display by color, in addition to displaying the congestion by the numerical value. By displaying in color, dispatchers can easily identify congestion.

First, we verified accuracy of the numerical value. We verified the error of estimation using MAPE, which is an indicator for evaluating magnitude of error. We found an error of 10% to 60%. In the calculation method of MAPE, errors become large as the number of people is small. With this level of error, we can use the tool for business use.

ABSTRACT. Australia’s first autonomous train began running in July 2018. Its running was preceded by extensive trials of both on- and off-train technology. It was not a classic metro train but a 30,000+ tonnes bulk iron ore train, comprising 220-240 wagons, each weighing 130-160 tonnes when laden, and hauled by 2x3280 kW diesel locomotives. This paper discusses the usual rationales for developing autonomous trains and then tests them against the realities of running heavy haul freight trains in remote areas. Any apparent lack of line capacity is less important than the need for reliable mine-to-port supply chains. Furthermore, mining in remote areas is expensive and increasingly difficult to resource so automation of processes is increasingly attractive to mining companies. The automation of iron ore railway operations beckoned if mining companies could assemble, test and have accepted the various technical building blocks. Pilbara Iron has now completed these steps.

ABSTRACT. The increasing movement of goods in Europe due to the globalization and the rapid rise of e-commerce is a huge challenge for the governments, which have to deal with congestion on the roads and environmental pollution. Using train for freight transportation can help to tackle both problems: because of its high loading capacity and because in Europe, railway networks are mainly electrified. However, freight transportation by rail in Europe is much reduced when compared to road transportation. As a consequence, the European authorities try to improve rail infrastructures and connectivity, to increase the efficiency of rail transport and also to increase its share in the European transport market to decongest roads and to reduce pollutant emissions. In this paper, a mathematical programming-based model is presented for assessing a capacity expansion problem on a railway network. In this model, the authors have considered extensions of elements of an existent freight railway network, jointly with actions on the network with relative smaller cost, such as the inclusion of new sidings or new gauges in several rail segments, expansion of classification terminals or stations, and also capacity enhancements by new blocking/control systems. These aspects are usually not taken into account in models for regional planning. Our approach, rather than a model of railway capacity expansion can be considered a mixture of capacity-expansion with network design. The model is tested on a small regional network of the Mediterranean Corridor, and the computational results show its applicability to larger networks.

ABSTRACT. To remain competitive in the competition on the cargo market, railway undertakings need to execute operations with high efficiency. IT systems can contribute to determine the actual boundaries of the use of available production resources and operate optimally close to these bounds. Mathematical optimization and automation are key factors to realize this, e.g. in the field of Locomotive rotation planning. This contribution describes different approaches that have been implemented in the Locomotive Optimization System LOOP which has been developed by DXC Technology in close cooperation with Green Cargo – the largest rail cargo operator in Sweden. The main purpose of the system is to create plans for template weeks of cyclic planning (e.g. for simulation studies on the procurement of locomotives of different type) and dated monthly plans for short term planning. Different mathematical models are used to consider the various aspects of the objective function, e.g. number of locomotives used, operational efficiency and penalisation of loco changes in multiple traction. The consideration of conditions imposed by surrounding planning problems (e.g. crew, wagon transport) has been one of the biggest challenges in the project. The resulting mixed integer linear programming models are solved by commercial solver (CPLEX). The planners work in a web-based user interface, which is based on DXCs Rail Cargo Management Solution used by more than 20 customers throughout Europe. The close cooperation between the railway operator and the IT solution provider resulted in a solution which allows building plans in shorter time than previously and saving locomotives.

ABSTRACT. Air and high-speed rail (AH) integration services are gaining ground with the development of the high-speed railway and the airline industry. A well-designed feeder train timetable with good transfer quality in the AH integration service is of great significance, especially when train frequencies are low in a transfer node. To assess the transfer quality, we classify the transfers based on the transfer time. In this study, a bi-objective train timetabling model is proposed to maximize the quality of transfers from trains to flights and minimize the deviation from the original official timetable in the AH integration service. By optimizing the two objectives independently in two stages, a heuristic algorithm is developed to solve the proposed model. In the first stage, the first objective is optimized; in the second stage, the timetable shift is minimized by the rolling horizon approach with the transfer quality fixed. The method is applied to Shijiazhuang Zhengding International Airport, China. The result indicates that the proposed model is effective in improving the transfer quality and decreasing the timetable shift, and some analysis of solution are presented to verify the efficiency of the suggested model.

ABSTRACT. In this article, we present a practical approach for the optimized creation of railway timetables. The algorithms are intended to be used by Deutsche Bahn, Germanys largest railway infrastructure provider. We show how our methods can be used, both for creating a timetable in advance and for answering ad-hoc requests coming in via a digital app. Numerical experiments are provided to show that our solution exceeds manual timetabling in terms of capacity usage, travel times and the time taken for creating the timetable.

ABSTRACT. There are many stakeholders when it comes to railway timetable planning, e.g. infrastructure managers, railway undertakings and train passengers. This paper analyses timetable quality from the perspective of the Swedish infrastructure manager, i.e. from the perspective of an infrastructure manager operating in a deregulated market. Seven categories of timetable quality are discussed: feasibility, disturbance resistance, competition management, capacity safeguarding, application fulfilment, attractiveness and compatibility with surrounding planning areas. Each category is introduced, including references to legal documents, current development projects and research literature. Further, an interview study with eight practitioners gives insight into the current state of practice in Sweden. The practitioners consider feasibility to be both most important and easiest to handle. Capacity guarding is considered least important, despite its prevalence in legal documents and envisioned process developments, and is also considered hardest to handle. The lack of published guidelines was repeatedly mentioned as an explanation to why capacity guarding is not considered during timetable construction. In general, formal rules and guidelines seem important for supporting the timetable planners in their arbitrating role, and improved timetable planning tools would also be beneficial for resolving e.g. problems with maintenance possession planning. The results show that there is a gap between the wanted state as depicted by legal documents and development projects, and the current state of practice in Sweden. Operational research can contribute to closing this gap, both by constructing formal guidelines and measurements for quality aspects, and by developing functionalities for timetable planning support tools.

ABSTRACT. On the network of the Swiss Federal Railways (SBB) there is huge variability in the energy consumption for comparable train runs. Consequently, there is a significant potential to achieve energy savings with improved driving strategy, which can be influenced by providing useful information to the train driver. As part of the smartrail programme operated by the Swiss railway industry, several energy savings measures are due to be implemented. As a first step in the smartrail energy measures, SBB conducted a pilot test in summer 2018. This pilot involved 473 test runs on two important passenger trains in Switzerland: the long-distance train IC5 and the local train S12 from Zurich. For each train run, based on effective routing, train composition, speed restrictions and timetable fixed points, a speed profile and new service times for each station were calculated early each morning for all the train runs of the day. More than 80% of the regular train drivers would welcome the rollout of the new timetable information soon. A comparison of the accompanied journeys against the ‘baseline’ (same trains in the same period) shows a significant reduction in energy consumption without affecting punctuality: depending on the train journey, the accompanied runs consumed between 1.4% and 13.3% less energy per gross tonne-kilometre. The high levels of acceptance by the train drivers combined with the significant energy savings achieved without affecting punctuality is very promising. For this reason, a system-wide rollout is currently being investigated and could be started by late 2019.

ABSTRACT. Catenary-free transportation system is in development and has been installed in several countries. Battery train is used as the train which can travel in catenary-free section using power supplied from an onboard storage system such as lithium-ion battery. However, as characteristics of this type of train, an energy consumption of battery train depends on the state of energy of the storage system. Furthermore, battery train needs rapid charging when running the long distance. Hence, energy efficient design of catenary-free transportation system is important. In consideration of these characteristics, in this research, we propose a generation method of the timetable including running time, dwell time and location of charging infrastructure which is the most energy-saving for catenary-free transportation system with battery trains. Firstly, we conduct a running simulation and reveal the relationship among running time, state of charge of the battery and energy consumption. The characteristic is derived as a two-variable function and nonlinear for each variable. Although this optimization problem can be defined as a nonlinear programming problem, we ease to solve this problem using linear approximation to the energy consumption characteristic. Specifically, we use a method of dividing the characteristic defined on the space as curved surfaces into fine lattice shapes and approximating it as a polyhedron composed of minute triangles. In the end, we carried out a simulation in a simple case so that we can show the effect of the proposed method.

ABSTRACT. Energy efficiency is an important topic for railway companies wishing to reduce CO2 emissions and save money. One of the research areas to improve the energy efficiency of railways is energy-efficient train control (EETC). EETC is an optimal control problem with the aim of finding the driving strategy or trajectory that meets the timetable with the least amount of energy consumption. The potential for EETC is determined by the timetable, because the running time supplements determine how much energy can be saved by energy-efficient driving. Therefore, in this paper we focus on the optimal distribution of running time supplements for the energy-efficient driving of a single train over multiple stops. Furthermore, we compare an indirect solution method to a direct solution method to solve the EETC problem, and apply the two methods to different case studies to find the optimal distribution of running time supplements to achieve the most energy-efficient driving. The indirect method is used by the Energymiser Driver Advice Systems. Direct solutions are found using the Radau Pseudospectral Method. The results of the direct solution method confirm the optimality conditions used by the indirect method and also confirm that the optimal cruising speed is the same over multiple sections.

ABSTRACT. Problems of dynamic origin-destination, hereafter OD, matrix estimation using smart card data can be modelled as entropy maximization problems and efficiently solved using solution techniques such as Lagrangian relaxation. In this paper, we briefly review the literature of OD-matrix estimation and the use of smart card data. We show how this problem can be modelled and solved with incomplete smart card data where the trip distribution incoming to stations is known but not the outgoing distribution. A large non-linear entropy maximization problem is solved using an iterative algorithm solution based on Lagrangian relaxation. The algorithm is tested using a case study from the commuter train service in the great Stockholm region. Even though there are no theoretical guarantees for the algorithm’s convergence, the results show that with the incomplete smart card data, the solution method converges and finds an estimate of the dynamic OD-matrix reflecting the reported aggregate statistics from the local operator.

ABSTRACT. The paper presents the characteristics of the departing passenger flow in different stations based on the real-record passenger flow data of Beijing-Guangzhou high speed railway, from January, 2010 to December, 2015. The passenger dataset is framed for the long short-term memory (LSTM) model, considering the expectation input format of LSTM layers and the characteristics of the data. The Keras model in Python is used to fit LSTM model with tuning and regulating all the parameters necessary in the model. Then the fitted LSTM model is applied to forecast the short-term departing passenger flow of Beijing-Guangzhou high speed railway. The influence of important parameters in the LSTM model on the prediction accuracy is analyzed, and the comparison with other representative passenger flow forecast models is conducted. The results show that the LSTM model can get the valid information in a long passenger flow time series and achieve a better performance than other models. The passenger flow prediction errors valued by MAPE are 7.36%, 7.33%, 8.03%, respectively for Chenzhou station, Hengyang station and Shaoguan station. The parameters in the LSTM model such as the number of neurons, the input historical data length and the size of output layer have a great influence on the prediction accuracy.

ABSTRACT. Railway operations in dense traffic areas are very sensitive even to small disturbances, and thus require careful planning and real-time management. Dwell times in stations are in particular subject to a high variability and are hard to predict; this is mostly due to the interactions between passengers and the system during the dwelling process. This paper proposes an approach for estimating the minimum dwell time knowing the numbers of alighting, boarding and on board passengers, using Automatic Vehicle Location and Automatic Passenger Counting data. Based on the knowledge of this value, a method for estimating the conditional distribution of dwell time given passenger flows is presented. Numerical experiments are carried out on two stations located inside the dense traffic area of Paris suburban network. The obtained results show a broad applicability of these methods, that hence seem very promising.

ABSTRACT. The impact of rail network improvements on capacity and rail traffic punctuality in socio-economic analyses currently lacks an established quantitative method. This paper presents a professional research conducted for the Finnish Transport Agency aimed at the development of methods to evaluate such impacts. Two methods are proposed. The first is aimed at capacity estimation, and is an adaptation of the UIC 406 method to the characteristics of the Finnish rail network. The second is a method based on mathematical regression that allows estimating delays on lines given a set of parameters describing their characteristics. The delay estimation method proposes two distinct formulas for single- and double-track lines. The proposed methods were studied both on Finnish actual data and on a UK scenario. The use of these methods enables network managers to evaluate both network saturation and the effect of investments on delays in a simple way. This study has been approved and adopted by the Finnish Transport Agency.

ABSTRACT. Capacity assessment of high-speed railway corridor is critical in tactical planning process because it is beneficial to unearth the potential capacity and improve the capacity utilization without new investment in construction. China’s high-speed railway corridor serves trains with high heterogeneity in different route, speed, and stopping plans. This paper first illustrates the necessity of assessing the corridor’s capacity as a whole without decomposition. Based on the concept of base train equivalent (BTE), two methods named “capacity occupancy equivalent (COE)” method and “demand adaptation equivalent (DAE)” method are developed to standardize different types of trains into an equivalent unit. The case study of Jing-Hu high-speed railway corridor demonstrates that the methodology is concise in capacity assessment, and the impact of the long-distance direct service on corridor capacity utilization is also calculated.

ABSTRACT. Preventing passenger and freight trains from crossing each other in double-track rail-way tunnels is a fire safety measure required by the German railway authority to prevent fatal accidents. The prohibition poses a restriction on infrastructure usage that has to be incorporated in rail traffic planning. While it has already been implemented in timetabling and simulation tools, its effects on line capacity in long-term strategic planning has not been investigated so far. This paper presents a method to incorporate restrictions on sim-ultaneous track usage in the blocking time calculation and minimum headway time esti-mation. The effects on line capacity are analysed quantitatively based on the STRELE approach, which is an analytical method for strategic long-term capacity planning cur-rently used by German railway infrastructure manager DB Netz AG. Results are validated by comparison to delay increase in microscopic simulation of train operations.

ABSTRACT. This paper tackles the real-time Railway Traffic Management Problem (rtRTMP) of finding an optimal choice for the train schedules and routes to reduce the delays of trains due to unavoidable conflicts. We present a new formulation of the rtRTMP. This new formulation is based on a previously proposed one that models railway traffic at a microscopic level with optional activities using a Constraint Based Scheduling (CBS) approach. To ease the modelling of optional activities, a new concept based on a tree data structure and a specific filtering algorithm was extended by the introduction of conditional time-interval variables in CP Optimizer library. The new formulation of the rtRTMP presented in this paper exploits the conditional time-interval variables. The formulation has been validated with experiments on a large set of instances. The experimental results demonstrate the effectiveness of this new CBS model and show good performance of the proposed approach compared with the state-of-the art RECIFE-MILP algorithm.

ABSTRACT. Disruptions in urban rail transit systems usually result in serious incidents because of the high density and the less flexibility. In this paper, we propose a novel mathematical model for handling a complete blockage of the double tracks for 5-10 minutes in the peak hours, e.g., lack of power at a station, no train can pass this area during the disruption. Under this operating scenario, train services may be delayed or cancelled, some rolling stocks may be short-turned at the intermediate stations with either single or double crossovers. To ensure the service quality provided to passengers, the backup rolling stocks inside the depot may also be put into the operation depending on the consequences of the disruptions. Thus, the number of rolling stocks in the depot is considered. We discuss the disruption management problem for urban rail transit systems at a macroscopic level. However, operational constraints for the turnaround operation of rolling stocks and for the rolling stock circulation are modelled. A mixed integer linear programming (MILP) model is proposed to minimize the train delays and the number of canceled train services as well as to ensure a regular service for passengers, while adhering to the departure and arrival constraints, headway constraints, turnaround constraints, service connection constraints, inventory constraints, and other relevant railway constraints. Existing MILP solvers, e.g. CPLEX, are adopted to compute near-optimal solutions. Numerical experiments are conducted based on real-world data generated for Beijing subway line 7 to evaluate the effectiveness and efficiency of the proposed model.

ABSTRACT. One of the crucial factors in achieving a high punctuality in railway traffic systems, is the ability to effectively reschedule the trains during disturbances. Railway rescheduling is a complex problem to solve both from a practical and a computational perspective. Problems of practically relevant sizes have typically a very large search space, making it a challenge to arrive at the best possible solution within the available computational time limit. Though competitive algorithmic approaches are a widespread topic of research, not much research has been done to explore the opportunities and challenges in parallelizing them on Graphics processing units (GPUs). This paper presents a conflict detection module for railway rescheduling, which performs its computations on the GPU. The aim of the module is to improve the speed of solution space navigation and thus the solution quality within the computational time limit. The implemented algorithm proved to be more than twice as fast as the sequential algorithm. We conclude that for the problem under consideration, using a GPU for conflict detection likely gives rise to better solutions at the end of the computational time limit.

ABSTRACT. We assess the validity of the fixed-speed approximation for train speed dynamics in the real time Railway Traffic Management Problem. This is done through a statistical analysis on a number of perturbed scenarios on different railway infrastructures, for different objective functions commonly used in the literature. For each scenario, we analyze the ranking of the generated solutions both in the fixed-speed approximation, obtained by solving the optimization model, and with the variable-speed dynamics, obtained through micro simulation with the OpenTrack software. Our results indicate that some objective functions are somewhat reliable when used in conjunction with the fixed-speed approximation, while others require more detailed studies.

ABSTRACT. Unexpected disruptions occur frequently in railway systems, during which many train services cannot run as scheduled. This paper deals with timetable rescheduling during such disruptions, particularly in the case where all tracks between two stations are blocked for a few hours. In practice, the disruption length is uncertain, and a disruption may become shorter or longer than predicted. Thus, it is necessary to take the uncertainty of the disruption duration into account. This paper formulates the robust timetable rescheduling as a rolling horizon two-stage stochastic programming problem in deterministic equivalent form. The random disruption duration is assumed to have a finite number of possible realizations, called scenarios, with given probabilities. Every time a prediction about the range of the disruption end time is updated, new scenarios are defined, and the model computes the optimal rescheduling solution for an extended control horizon, which is robust to all these scenarios. Based on the model, uncertain disruptions can be handled with robust solutions in a dynamic environment. The stochastic method was tested on a part of the Dutch railways, and compared to a deterministic rolling-horizon method. The results showed that compared to the deterministic method, the stochastic method is more likely to generate better rescheduling solutions for uncertain disruptions by less train cancellations and/or delays, while the solution robustness can be affected by the predicted range regarding the disruption end time.

ABSTRACT. Due to high demand and high capacity consumption, railway timetables often become sensitive for disturbances and there is little time in the timetables for delay recovery. To maintain a high quality in railway traffic it is important that the timetables are robust and there is a need for strategies and rules for how to make them robust without consuming too much capacity. In this paper we present how timetable rules can be implemented manually to create more robust timetables. The rules are separated into two categories, rules to make the timetable feasible and rules to increase the delay resistance and recovery. The implementation is illustrated in a real-world case from when the timetable for the Swedish Southern mainline was created for 2019. In the paper we describe how new rules can be applied manually and discuss advantages and disadvantages by using this approach. We also describe how the rules effect the trains, their timetable slots and runtimes. The results from this study show some of the difficulties when moving from theory to practice and what can be done with limited resources in reality. It gives insights to the practical approach of train timetabling problem which can be used to improve optimization models.

ABSTRACT. This paper deals with robustness evaluation at station, and in particular for the train platforming problem (TPP). This problem consists in a platform and route assignment in station for each scheduled train. A classical robustness evaluation is simulation: simulated delays are injected on arriving and departing trains then propagated, and results are averaged on a large number of trials. A robust solution of the TPP aims to limit the average amount of secondary delays. However, a simulation framework at station is difficult to calibrate : it requires a realistic delays generator and an accurate operating rules modeling.

This paper proposes an original simulation framework using classical statistical learning algorithms and calibration assessment methods to model simulation inputs. This methodology is applied on delay data to simulate delay propagation at station. It highlights the importance of delay calibration by showing that even slight miscalibration of inputs can lead to strong deviations in propagation results.

ABSTRACT. Railway operation companies provide a more efficient and sustainable service for passengers, so that they can have a stronger competitiveness in the multimodal transportation market. However, in daily railway operation, inevitably unplanned events occur several times per year such as rolling stock break-down, which may influence train running time, as well as arrival and departure time. Under severe disruptions, stop patterns of trains may be changed, even cancelling or inserting certain trains may be taken by dispatchers. Train rescheduling will be quite different and challenging in a railway system with ticket booking mechanism compared with non-reserved mechanism. This paper develops an mixed-integer programing model for the problem of train rescheduling with passenger reassignment on a railway network with ticket booking mechanism under severe disruptions. Passenger reassignment will be taken into consideration to ensure that as many as passengers effected by disruptions may arrive at their destinations as early as possible. The function objective is to maximize transported passengers of cancelled trains, and minimize total delay time of trains at their destination stations, with consideration of planning extra stops for unaffected trains to transport more passengers effected to their planned destinations, seat capacity limitation and uncertainty of disruptions. A constraint will be set to ensure that the same number of passengers are assigned to the same following train(s) under different random disruption scenarios, which imposes the robustness of dispatching. There will be several numerical experiments based on “Beijing-Shanghai” High Speed Railway Line to demonstrate the validity and efficiency of our model.

ABSTRACT. The Chinese high-speed rail network has a fast-growing number of electric multiple units (EMUs) in service and is facing increasing pressure of maintaining all EMUs on-time. The capacity at an EMUs depot is relevant to its track utilization rate, which can be improved by a better shunting operation plan. An EMUs depot typically consists of a maintenance yard, washing yard and temporary storage yard. Each track in those yard has two sections, and can be occupied by a long EMU or two separate short EMUs. The two yard types, stub-end and through, further add complexity to the shunting operation problem. Compared to the previous researches, the shunting operation plan studied in this paper takes the yard types and the section assignment into account simultaneously. An optimization model was established aims to minimize the total delay time of EMU in running shed during the plan horizon. The constraints include the numbers of operation tracks and EMUs, operation sequence and the dwell time of operation tracks, etc. The original problem is transformed into a typical job shop scheduling problem with additional space and time constraints. Then a hybrid heuristic algorithm based on Tube Search is designed to solve the model. Finally, by taking a real-world EMU depot as an example, the numerical results show that the proposed solution method can yield an assignment plan with the optimal track utilization in a small amount of computational time and can be implemented in a computer-aided planning system easily.

ABSTRACT. Intermodal transportation systems play an important role in fulfilling the growing market needs for freight transportation. Integrity of operations in such systems can be obtained by real-time, and just in-time communication among the involved actors. In In2Rail and X2Rail2 European projects, an integrated communication platform (the integration layer) is developed as a communication medium between railway services, applications, and external systems. This paper introduces the integration layer and shows how this platform can be deployed to enhance the operations of intermodal freight transportation, with a special focus on the management of dynamic demand.

ABSTRACT. Single-railcar shipments of freight that move in multiple freight trains and are sorted at several classification (marshalling) yards during their trip from origin to destination remain an important source of traffic and revenue for North American freight railways. The train plan for a carload freight railway network determines how railcars are sorted into blocks by common destination and transported on trains between yards. To deliver competitive service to freight shippers, practitioners must devise an optimal plan that balances mainline and yard efficiencies while dealing with variation in both train arrival times and inbound traffic volumes. Despite the important role of both mainlines and yard facilities in freight rail transportation performance, little attention is devoted to investigations of classification yard performance, and there are few yard capacity models and tools. To address this railway industry need, the research in this paper seeks to investigate the influence of inbound traffic volume variation and schedule flexibility on classification yard performance and capacity. A series of simulation experiments quantify the interaction between arrival time and volume variability as measured by different yard performance metrics. The simulations are conducted with a discrete-event simulation model developed specifically for analysis of hump classification yards. Preliminary results suggest that increasing schedule flexibility causes classification yard performance to decline. Increasing volume variability appears to have a less pronounced effect. The results of the research will allow railroads to make more informed business decisions regarding train operating plans and make more efficient and economical use of existing yard capacity.

ABSTRACT. Single-railcar shipments of freight that move in multiple freight trains and are sorted at several classification (marshalling) yards during their trip from origin to destination remain an important source of traffic and revenue for North American freight railways. Much of this freight moves in trains that depart yards when a certain number of railcars are ready to be moved, and not according to a pre-planned timetable. Recent North American industry trends have seen a move away from these flexible operations to more structured operations where trains depart yards at specific times according to a train plan. This research investigates how operating strategies at classification yards and schedule flexibility of mainline trains combine to affect the average railcar transit time from origin to destination across a representative rail network. Experiments comparing different yard operating strategies under varying degrees of schedule flexibility were conducted using SIGMA simulation software. Simulation results suggest that railcar transit time increases as the level of schedule flexibility increases under all three of the studied operating strategies. Several factors, including the timing and distribution of railcars on inbound trains arriving at the yard, and frequency of outbound train departures, influence the sensitivity of transit time to schedule flexibility under different operating strategies. There is no universal best strategy to minimize railcar transit times; different traffic scenarios require different operating solutions. This research may help railway practitioners develop more effective operating strategies to improve carload freight operations and performance, and influence decisions on yard operations and train assembly planning.

ABSTRACT. This paper presents a review of research and models regarding sustainability of railway passenger services. In order to take into account all relevant aspects in terms of environmental impacts of a railway passenger service, a holistic system perspective is required, that includes a whole life cycle assessment. A life cycle approach is important since comparison of for instance only the exhaust emissions of an electric vehicle with a petrol vehicle is misleading, due to neglecting the emissions of for instance electrical energy production process. Thus, all stages in energy carrier, vehicle and infrastructure life cycles are to be considered. Existing models are analyzed, as well as possible developments, focusing on diesel and electrical traction as the most common traction options in use, and on GHG emissions, especially on CO2, which takes the greatest part in all emissions. Issues and challenges in improving the environmental impact of railway passenger services are addressed. Additionally, several areas are indicated where environmental aspects could be included in future assessment models. The main challenge is answering how the existing partial assessments can be brought together and, together with filling the identified gaps, allow to conduct a comprehensive LCA which will produce real-world emissions estimations. Results of this paper will be used as an input in developing a framework for quantifying and improving overall environmental impacts of a railway passenger service.

ABSTRACT. Seat inventory control is a railway revenue management method that aims to maximize profits by determining the availability of products via a prepared rule/strategy (referred to as the seat inventory control policy). In past decades, the most famous type of seat inventory control policy was partitioned booking limit control (PBLC), which was extensively applied by railway companies but observed to be inefficient for stochastic demand. In recent years, China Railway Corporation attempted to overcome this deficiency by applying a new type of seat inventory control policy, denoted “seat-based control”. Seat-based control provides a more flexible way to manage resources but has encountered difficulty in setting the initialization parameters. This study focuses on the parameter setting problem for seat-based control considering i) the randomness of customer arrival, ii) the customer choice behavior and iii) the specifics of China Railway Corporation. A Markovian decision process (MDP) model is built, and a genetic algorithm method with a special structure is designed. The performance of the seat-based control is tested in two experiments with two other benchmarks. Finally, we apply our method to practical data of the Nanning-Guangzhou high-speed railway line.

ABSTRACT. Passenger and freight traffic growth on Britain’s railways has led to increased needs for maintenance, renewal and enhancement of the national railway network, and reduced opportunities for access to the network to conduct these engineering activities without disrupting operations. As a result, the costs of compensation to operators for service disruption and revenue loss have been increasing in line with traffic levels. There tends to be a trade-off between the cost efficiency of engineering activities and the compensation costs for the operational disruption caused, since longer track possessions are typically more efficient, but also more disruptive, reducing network availability for operations. There is thus a need to reduce and, ideally, minimise the total costs of engineering activities and compensation for the disruption caused. The current possession planning process does not actively aim to minimise service disruption and compensation costs, much less the combined engineering and compensation costs. A further, detailed review of the current possession planning process, including data availability and needs, is being undertaken, and, together with the results of recent research, will be applied to (i) amend the current possession planning process to reduce its disruptive impact and compensation costs, thus increasing network availability for operations, and (ii) to identify data requirements to enable the assessment of duration, engineering costs and timetable impacts/compensation costs associated with alternative possession strategies, and apply these in combination with scheduling techniques to reduce and, ideally, minimise combined engineering and compensation costs.