We are pleased to announce that the Autumn 2024 Viriato Standard and Enterprise releases have been delivered to our customers.

In addition to the ongoing performance and stability improvements we continuously undertake to improve Viriato, and the implementation of customer specific functionality delivered in their own program versions, the following features are some specific highlights of this product release available to all:

• This release switches to the Microsoft .NET 8 framework. This provides the long-term assurance that Viriato is maintainable for the future.
• To improve security, all program components are now digitally signed with a digital certificate to guarantee the authenticity of the software.
• The railML 2.2 importer handles defined remarks in the train path and timetable nodes, and the route finding has been improved for trains without section information allowing a choice of routing strategies.

• Microscopy on Demand has been extended to include new user selectable behaviour when the microscopic and Viriato models have differences. Also, when working with different infrastructure variants in Viriato Enterprise the conflicts and block occupations indicate whether they belong to the same infrastructure or not. MoD is also now fully functional in Viriato Enterprise and can be linked with different microscopic models for alternative infrastructures.
• In the robustness analysis the primary delay evolution of a robustness analysis run can now be exported to a spreadsheet with the existing outputs.
• A new platform assignment funder module has been added which can search for conflict free platform assignments in stations for existing timetables. This optional module also requires a separate Gurobi licence.

If you are interested in learning more about this release, or Viriato in general, please do not hesitate to contact viriato@sma-partner.com.

A further step towards automatic conflict resolution with MoD: Automatic changing of routes in stations. Congratulations to Jonathan Gut for successfully defending his thesis in which he extended existing work to resolve conflicts through re-routing.

SMA would like to thank his supervisors at TU Dresden Prof. Dr. Karl Nachtigall and M.Sc. Maik Schälicke for the excellent #openviriato collaboration.

In both medium-term and long-term timetabling, it is generally the case that not all microscopic infrastructure data is available.

The advantages of a macroscopic or mesoscopic infrastructure model when compared to a microscopic model are well understood for these planning phases and have been demonstrated in many projects. However, there are important aspects in medium-term timetabling that deal with the operational feasibility of a timetable and where conflict detection between trains is needed. In some circumstances, conflict detection based on general separation times may be adequate. However, in cases with dense traffic on complex topologies, the results from such an approach are generally not sufficient. In such cases it would be preferable to use microscopic conflict detection in medium-term timetabling.

In this article, we present advancements in Microscopy-on-Demand (MoD) that allow microscopic calculations to be performed in medium-term planning. We call this functionality “Partial Coverage”. By this we emphasise that microscopic infrastructure is only available for a partial area of the network, or that the infrastructure in the macroscopic client and the microscopic server do not exactly match.

We are pleased to announce that the Spring 2024 Viriato Standard and Enterprise releases have been delivered to our customers.

• In addition to the ongoing performance and stability improvements we continuously undertake to improve Viriato, and the implementation of customer specific functionality delivered in their own program versions, the following features are some specific highlights of this product release available to all:
• Timetable remarks can now apply to the whole train and be overridden for individual portions of the train run.
• The Viriato Enterprise mini calendar has now be reorganised so that the functions take up less screen space to improve usability.
• TAF/TAP-TSI fixed nodes can now be defined, and a plausibility check is available to ensure that the user has complied with the rules that trains should be planned to stop at these nodes.
• TAF/TAP-TSI lead RU’s can be updated for multiple trains in bulk.
• TAF/TAP-TSI relevant information has been added to the railML export.
• When importing trains using railML and the data contains differences between the infrastructure in the railML and the Viriato database, the user now has more options for how to treat the missing nodes to ensure that the imported trains are valid in Viriato.
• Rolling stock definitions now have a dated validity, allowing improved management of rolling stock by creating multiple versions of a vehicle which differ in certain attributes.
• In the customer timetable it is now possible to see which operational trains form a commercial train.
• In the graphic timetable and platform occupation views, the line number of a train can now be displayed as a label.
• In the platform occupation view when viewing a topology, the direction of train movements has been clarified with arrows, and the ability to select trains which overlap on the diagram has been improved.

• In Microscopy on Demand (MoD), the use of mixed models where part of the infrastructure is mapped to a microscopic model in an external system, and the remainder of the infrastructure is only mesoscopic has been implemented. This enables the use of larger models and the focusing of detailed effort on areas of specific interest.
• In MoD, the display of signals and stopping points has been added to the topology viewer.
• In the trip time analysis module, when working with commercial trains it is now possible to open these trains from the analysis view.
• The running time calculator now has a user configurable default speed for the case where the speed profiles are not fully defined.
• The robustness analysis can now be carried out only for a specific region of the Viriato infrastructure, allowing the user to focus on the area they are studying without having to model the entire network with the consequential performance implications.
• The path search module now includes additional settings allowing finer control over how the conflict free paths are found. These include settings which control how much additional time can be added to a train run, whether reserves should be kept for performance purposes, the ability to ignore capacity restrictions in nodes where the user believes sufficient capacity will be available regardless of the analysis, an problem analysis mode which identifies the nodes with insufficient capacity when conflict free paths cannot be found and the improved selection of multiple trains for the case where the user is inserting different trains sequentially into the timetable.
• The train conflict report in the conflict detection now considers infrastructure variants and versions when working in Viriato Enterprise.

If you are interested in learning more about this release, or Viriato in general, please do not hesitate to contact viriato@sma-partner.com.

Our initiative #openviriato has led to another successful collaboration: Together with supervisors Prof. Dr. Karl Nachtigall, M. Sc. Maik Schälicke and the student Jonathan Gut from TU Dresden we extended our automatic conflict resolution algorithm to redistribute the running time reserves of trains using MoD services. Jonathan's important contribution and implementation showed that the chosen approach is a very promising step towards real-life automatic microscopic conflict resolution.

Because of these excellent results we are going to pursue this topic further. Expect to hear more about this soon. We thank everyone involved for the good cooperation with Jonathan's student thesis (Studienarbeit). Congratulations to Jonathan!

The Viriato timetable planning system has now been in use at DB Regio AG and DB Fernverkehr AG for around 25 years. The contract for the maintenance and further development of Viriato Enterprise has recently been extended for a further eight years.

Viriato is currently used by both companies for different planning timeframes (multi-year planning, annual planning and sub-yearly planning). The core task consists of the train path ordering process for the annual timetable at DB InfraGO AG and the distribution of contracted train paths to other processes.

SBB AG has been using Viriato for long-term and medium-term planning since 2003. The maintenance contract for Viriato Enterprise has now been extended for a further four years. In long-term planning, SBB uses Viriato for network design and to develop the formal expansion stages and network utilisation concepts (“Netznutzungskonzepte” or “NNK”) on behalf of the Swiss Federal Office of Transport (FOT).

For medium-term planning, Viriato is used to develop concepts for construction works and the network utilisation plans (“Netznutzungspläne” or “NNP”), again on behalf of the FOT. The latest NNP approved by the FOT serves as the basis for the TTR Capacity Strategy for the corresponding year.

sma.software has introduced the label #openviriato to identify collaborative projects based on our Viriato software between industry and academia. As we explained previously, with this type of collaboration we want to contribute to accelerating urgently needed digitalisation efforts in the capacity management role of railway Infrastructure Managers (IMs).

Within the scope of an internship with Luca Bataillard, a student of computer science at EPF Lausanne, three different aspects of this collaboration concept could be implemented simultaneously.

The first is the collaboration with the EPFL as a representative of the academic world and through the topical context of the Microscopy on Demand (MoD) concept the second, in the further development of the integration of a third-party software from our industrial partner VIA-Con.

Finally, the third aspect of our strategy was aligned with the later part of the internship through the inclusion of the Viriato algorithm platform. This is aimed, among other things towards the accelerated development of optimisation and automation algorithms for capacity management.

DB Netz's capacity management is increasingly relying on Viriato and Microscopy on Demand (MoD) to construct the Medium-term concept for optimized capacity utilization (mKoK).

In the international timetabling process, the mKoK is to provide the basis for the TTR Capacity Model. In addition, DB Netz is using the mKoK to develop key process elements that are essential for implementing the Deutschlandtakt.

In this post we want to outline the high-level architecture of Viriato's macroscopic train simulator, which will be used for Viriato's new robustness analysis module currently being implemented by our development team. Roughly speaking robustness can be seen as the ability of a timetable to recover from delays caused by unforeseen events, including the potential rescheduling efforts. For a more precise definition we refer the interested reader to our annual report 2021 (pp. 18-20). In a Viriato robustness analysis we investigate the effects on the planned timetable, and how long it takes to return to normal train operation, after a disturbance has occurred on the network. 

The core of our robustness tool is a macroscopic train simulator. For a background about the relationship between train simulation and robustness we recommend our previous posts 04.11.2020 Robustness vs. Train Simulation and 11.06.2021 - Robustness Analysis through the Algorithm Platform.

As part of our open collaboration initiative #openviriato with industry and academia, sma.software is helping ETH Zürich’s IVT in their project to develop a method for calculating the total energy consumption by all trains in a railway network.

We have provided a research licence of our ZLR running time calculation service to the “RailPower” study jointly undertaken by the IVT and SBB.

We are happy to help contribute to the success of this project.

SMA would like to congratulate our #openviriato research partner Nikola Obrenovic and his team from the BioSense Institute for winning the #CopernicusMasters #SmartMobility Challenge. Nikola and his team proposed using the Copernicus Emergency Management Service for disruption management based on satellite data: Through image processing on the satellite data, disruptions on a railway network can be detected and using an algorithm that they are currently working on, reschedule the trains originally sourced from a timetable planned in Viriato in order to return the perturbed situation to a workable timetable.

The timetable data can be obtained using the Algorithm Platform, and the calculated result written back to Viriato where it can be visualised and used for further analysis.

The Algorithm Platform has been mentioned in the keynote talk of Dr. Thomas Schlechte from LBW Optimization at the RailBeijing 2021 International Conference on Railway Operations Modelling and Analysis (ICROMA).

We would like to thank LBW Optimization for their collaboration in #openviriato by choosing to use the Algorithm Platform as a repository and visualisation tool for their algorithm "raillation".

An implementation of an algorithm based on SPOT using Viriato's Algorithm Platform.

Abstract

SPOT [1] is a mathematical model for strategic passenger railway planning building on the well-known PESP (Periodic Event Scheduling Problem [2]). The goal of the SPOT model is to obtain an automatically generated and workable timetable during the strategic planning phase as it aims to provide a passenger-centric timetable.

We want to provide an implementation based on SPOT using Viriato's Algorithm Platform to deliver a software prototype that can be actually used by a subject-matter expert in practice so that the model's results can be assessed by them without any mathematical or programming background. We demonstrate the benefits that come with our Algorithm Platform to the researcher.

Our Goals for this Implementation

We want to highlight the benefits that come with our Algorithm Platform to the researcher:

• Data Acquisition and Provision. The Algorithm Platform retrieves all data requested by the algorithm from Viriato's database and provides it via an interoperable REST interface. There is no need to write database queries.
• Rapid Development. The input data provision and the simple way of passing parameters in combination with the predefined domain data model (AIDM) reduce the development effort considerably.
• Prevention from Misuse. Relying on the Algorithm Platform reduces the chance for the algorithm developer to make errors, and also protects them from erroneous data due to the enforced invariants in Algorithm Platform's Algorithm Interface.
• Visualisation of Results and Reports. The user can easily explore the solution which was written back to Viriato, allowing them to inspect the structure of the results in the available modules and assess their correctness and quality. In addition, reports in form of Excel sheets are generated to present the parameters used and a summary of the solution to the user giving them insights.

Note that our implementation of SPOT deliberately deviates in some aspects from the original model in [1] in order to enhance the applicability in practice. The main goal was to demonstrate the use of the Viriato Algorithm Platform rather than an analysis of the model.

SMA would like to congratulate Benoit Pahud on the successful completion of his MSc thesis at the Ecole polytechnique fédérale de Lausanne (EPFL) with industrial support from SMA. Benoit worked on a railway timetable rescheduling problem for disruption management with constraints on the passenger seating capacity. He brought theoretical research into practice through the implementation of an algorithm from the scientific literature using #openviriato. 

His work extended the existing model by adding the vehicle passenger capacities, with a consequential increase in the realism of the results.  We would like to offer a special thank-you to Prof. Michel Bierlaire, Dr. Nourelhouda Dougui, Stefano Bortolomiol and Marija Kukic from the EPFL Transport and Mobility Laboratory for their excellent collaboration on the project. Benoit’s work is available on GitHub.

We would like to present our robustness analysis tool, which we have applied in the context of a case study that we have carried out for a customer. Parts of it (e.g. the dispatching strategy) are connected to Viriato via the Algorithm Platform. We will show selected features of our tool, highlight the advantages of our approach using an external algorithm over a monolithic and closed implementation, and explain in which aspects a macroscopic robustness analysis can outperform a microscopic one.

 We will describe how the tool was helpful to us in the case study.

Since autumn 2020, the Timetable Concept Consulting and Market Launch Management department of DB Netz AG has been using Microscopy on Demand (MoD) productively. This enables their employees during early planning phases to carry out partial microscopic level analyses with reduced effort compared to previously. The integration of both micro- and macroscopic modelling worlds has increased their efficiency and planning quality. This is because the integration between macroscopic long-term planning and short-term microscopic planning avoids awkward switches between separate systems for the user, and reduces or even eliminates the number of iterations through both the microscopic and macroscopic worlds. This frees up valuable time for the user that can be used for productive planning tasks.

MoD combines the strengths of microscopic and macroscopic modelling without having to accept the disadvantages of them. At the macroscopic level, the infrastructure can be modelled with little effort and a largely conflict-free timetable can be quickly created for a larger network. However, operational characteristics such as routes with long separation times or overlap conflicts in the stations cannot be easily detected at a macroscopic level and often require microscopic analysis.

Our video illustrates this situation with two examples from Eiderbrücke and Husum on the Hamburg - Westerland line. In the illustrated timetable, no conflicts are initially discernible at the macroscopic level. However, the microscopic conflict detection shows problems in the two stations.

The detailed view in the expander provides initial information on the type and duration of the conflicts. By opening the Topo Viewer in Viriato, the user can see that at the Eiderbrücke station there is a short single track within the station, which cannot be depicted in the macroscopic model. The conflict there can be solved by creating sufficient separation times between the oncoming trains. A comparable situation exists at the Husum depot. In addition, the route of the train shown is not yet fully set, which can be remedied directly using the Topo Viewer. From the 2021 Viriato spring release, the Topo Viewer can also be opened directly from the conflict detection in a graphic timetable.

As part of a proof of concept project conducted for RFC North Sea – Mediterranean, one of the European Rail Freight Corridors, SMA has developed automated functions to evaluate capacity consumption and the residual capacity of international timetables considering both trains and temporary capacity restrictions (TCRs) for engineering works. Two algorithms have been used during this project with the Viriato Algorithm Platform: The first one determines capacity consumption by compressing the timetable and TCRs on each homogeneous section, junction and in stations. The second algorithm searches for available paths satisfying given constraints and performance goals in the input timetable and TCRs. Both algorithms have been tested on a complex and realistic example of an international network running from Antwerp (Belgium) to Saint-Louis (France). To the best of our knowledge, the algorithmic approach is novel and produces a broad variety of KPIs more efficiently than it would be possible to do manually, supporting the analyst by freeing them from monotonous work.

"Business Intelligence is the process of collecting, analysing and effectively presenting business data in order to make informed decisions. The RFC North Sea - Med handles capacity data, so why not develop a Capacity Intelligence approach? SMA perfectly understood our ambition, and their team combining timetabling, software and algorithmic expertise had the profile to match our expectations. The results of a Proof of Concept for Antwerp - Basel, one of our main routes, were extremely powerful and open up exciting opportunities for the creation of an international database of train paths and works, as well as for the visualisation of factual and objective capacity KPIs." 

Yann Le Floch, Managing Director, Rail Freight Corridor North Sea - Mediterranean

Microscopy on Demand in production: Last week, SMA has delivered its first Microscopy on Demand Add-on modules for productive use. Microscopy on Demand (MoD) refers to a conceptual software architecture that SMA has specified and developed over the last few years. 

This architecture makes it possible to integrate a microscopic infrastructure model with the macroscopic planning model for the specific tasks where microscopy is needed: i.e. running time calculation and conflict detection.

We are working on a macroscopic train simulator using our Viriato Algorithm Platform that allows us to apply customised conflict resolution strategies for robustness studies in Viriato. The goal of SMA's ongoing development is to validate that robustness studies can be carried out using an interaction of a dispatcher with a macroscopic traffic simulation of a railway network. We are proceeding with this by using the Viriato Algorithm Platform as a source for infrastructure and timetable data. Following this step we are going to enhance the simulator with features - e.g. to sample randomly stops on demand - to increase its analysis capabilities. From the study of the realisations of a timetable under perturbation scenarios, recommendations on how to plan robust timetables can be derived for train planners. We will also develop the framework to carry out Monte-Carlo types of analyses to give the user even more tools for robustness studies of timetables. The dispatching strategies are exchangeable in our model. 

Therefore, customers can also implement their own dispatching strategies using the Viriato Algorithm Platform to model the dispatching behaviour in their own networks and to study their timetables in cases of deviation from planned. Moreover, the results of robustness studies carried out by different dispatching strategies can be compared, which is interesting in its own right as potential changes in dispatching recommendations can be investigated. In this post, we are going to first explain the existing Viriato robustness module and then show its relation to a macroscopic train simulation. In a later post we will present the current architectural draft of this prototype.

Congratulations to Jordi Zomer for completing his MSc thesis at the TU Delft and SMA under the supervision of Prof. Rob Goverde and Dr. Nikola Bešinović. Well done! Jordi worked on an optimization model to (re-)schedule maintenance activities in an automated way using Viriato’s algorithmic interface for vehicle rostering. 

SMA thanks Jordi as well as our partners from TU Delft and Nederlandse Spoorwegen (NS) for this successful cooperation!

Using the Viriato Algorithm Platform, SMA is developing an automated function for rescheduling trains from a given timetable in the case of track possessions. This functionality has now been used productively for the first time in a consultancy project in Belgium. The following example shows an excerpt from a graphic timetable (also known as a train graph, time distance diagram or string line chart) in Viriato. In this case, the horizontal axis represents the timescale, and the vertical axis the distance with the stations and junctions along the infrastructure indicated by markers. Lines plotted on the graph represent the individual trains, with trains travelling in the forward direction over the infrastructure being plotted diagonally from top-left to bottom right, and trains travelling in the reverse direction being plotted from bottom-left to top-right. 

Stops can be observed on this graph where the plotted line is horizontal, i.e. for a period of time the train does not move over the infrastructure. In our example, the region with the pink background colour represents a section track closure, and the Viriato tooltip indicates that Track 1 is closed. This information is sufficient to enable an experienced train planner to identify the trains that have to be rescheduled due to the engineering works, but the algorithm detects this automatically for defining the search space.

SMA is working on a Python wrapper for the Algorithm Platform API, and are already collecting feedback from our beta testers in academia. Using the PyClient, it will be even easier than today to use Viriato programmatically in order to exchange data in algorithmic use cases.

Python’s intelligent code completion becomes available and facilitates the quick and Pythonic development of prototypes, in addition to the existing REST API. In just a few steps the results of the algorithms can be written back to Viriato and visualized.