Simulation
The simulation process is described in detail in the trenissimo manual.
Further links:
- Stochastic simulation: https://trenolab.atlassian.net/wiki/spaces/TRENISSIMO/pages/1049788/Export+and+Run+a+stochastic+simulation
- Scripted simulations: https://trenolab.atlassian.net/wiki/spaces/TRENISSIMO/pages/1049829/Scripted+simulations+and+output+scripts+advanced+feature
- Plugins in trenissimo: https://trenolab.atlassian.net/wiki/spaces/TRENISSIMO/pages/225280005/Additional+trenissimo+features+and+outputs
- Outputs in trenissimo: https://trenolab.atlassian.net/wiki/spaces/TRENISSIMO/pages/1050081/Outputs+of+trenissimo
In trenoLive the simulation setup allows the definition of input parameters for the simulation, such as stochastic inputs, train bundeling, rush hours and more.
Simulation area
The simulation area defines the geographic area where the simulation is conducted. A new area can be created by holding SHIFT and drawing a free-form area on the map with the mouse. An alternative method is to press ALT and click for straight lines from one corner to the next. The simulation area has to be an enclosed space. By drawing only straight lines the possibility to edit the simulation area after creation is greatly increased. A list of OCPs shows which stations are contained in the simulation area.
Rush hours
Rush hours can be used to define different stochastic inputs for certain periods in the simulation, for example on peak dwell time distributions. Rush hours are not obligatory for the simulation, if they are not defined all trains will be counted as "off peak".
Minimum stop times
Defining minimum stop times (MST) is also not obligatory. If it is done the simulation will take the inserted value as the minimum stop time. If it is not defined, the MST column in the train editor is exported. More information on the effect of MST is described here: Minimum Stop Times and Define the stochastic variables
Distributions
For the creation process of distributions see the dedicated page. The distribution curves define the stochastic input for stochastic simulations.
Simulation setup
The simulation setup is divided into three subpanels:
- Sim setup general data
- Train bundles
- Assignment of distributions to bundles
Sim setup general data
In the sim setup general data on the top left of the page it is necessary to insert name, simulation area and time period of the simulation. It is optional to add a description and to change the performance parameters. The performance parameters acceleration, braking and cruising percentages are deterministic values and are overridden in the stochastic simulation (see assignment chapter). The adhesion value is also deterministic, but can influence the stochastic simulation as well. Don't forget to save the setup. It saves also bundles and stochastic assignments, if present.
Train bundles
Bundling Trains means grouping all trains that will be simulated in a limited number of groups. The bundles are managed in the top right part of the page.
Bundling helps to:
To keep together all trains that have the same stochastic behavior (Reduce the effort when preparing a stochastic sim) To analyse the results of all similar trains together To apply current data to other timetables To manage the input in sensitivity analysis with many simulations
Trains can be bundled in three ways:
- In a very flexible way using the Bundles Creator criteria similar to the train searches.
- Based on single train codes
- (coming soon) Automatic train bundles by Category and Origin-Destination.
Bundles need a folder classification and a name. In the automatic bundle creation process, the folder is the category, the name is the Origin-Destination pairing. For each pairing a new bundle is created. Bundels can be merged after the creation. In the manual bundle creation process folders can be created without restrictions. If criteria based bundle creation is used, trains that match more than one criteria definition, will be bundled in the first one (first as from top to bottom). Bundels can be reorderd with drag and drop in the bundles list.
Assignment of distributions
Distributions are assigned for on peak and off peak periods or for the full simulation interval. A distribution is assigned to each bundle for each of the 4 input delay types:
- intial delay
- dwell time delay
- departure inaccuracy
- connection delay
The initial delay is assigned for on-peak and off-peak periods. The initial delay is applied to the train entering the simulation area.
The dwell time delay is assigned for on-peak and off-peak periods, too, and further divided into station load factors. The dwell time delay is applied to each train stop planned in the timetable.
The departure inaccuracy is applied to each train stop planned in the timetable, but only becomes relevant if the sum of MST and dwell time delay is smaller than the planned stop time. It applies a departure inaccuracy for trains, that would otherwise depart punctual on the second.
The connection delay is assigned for on-peak and off-peak periods. It is applied to trains with a connection inside the simulation area and only relevant for the outgoing trains (after the connection). The connection delay is optional. If no connection delay is assigned, the initial delay curve will be assigned for conections, too.
Simulation import into trenissimo
The completed simulation setup needs to be transferred to trenissimo. To prepare for a simulation in trenissimo a sychronisation between the micro and the macro model is advised to prevent import errors due to mismatches between the models. If you are sure that the models are synchronised, you can skip this step.
Open the trenissimo model. Open the properties of the regulation subproject and insert the web adress of your trenolive server installation. It is tipically composed like this:
https://asdf.live.trenolab.com
Select the project, if multiple projects are on the same server and if necessary select the correct registers for Short Code and Display Name. Click OK.
After that, go to the Macro folder and right-click on it to open the menu. Select Import from Treno. The trenoLive macro model will be compared to the microscopic model of trenissimo. If errors appear and the build failed the inconsistencies between the macro model and trenissimo should be redeemed.
If the build is successfull, the timetable and simulation setup can be imported. Go to the Timetable folder and right-click on it to open the menu. Select Import from Treno again. A dialogue window opens to select the trenoLive scenario within the project connected in the Regulations. If necessary select the correct registers for Short Code and Display Name again.
In the last step select the Timetable with the corresponding simulation setup created in trenoLive. Only the sim setups of the selected scenario are shown here. Click Finish and the timetable gets imported. Only the trains contained in the sim setup are imported. Depending on the characteristics of the simulation setup (deterministic or stochastic) delay distributions are imported or not. The import finishes with error or warning messages, which should be checked for problems with the timetable. The meanings of some error messagges can be found here: table at the bottom of the page
If the import does not give critical errors or warnings, execute the build of the simulation or the stochastic simulation subproject. Make sure that the subprojects are correctly linked in the Properties, especially when working with various subfolders of the same type in one trenissimo project. To update the simulation settings use the Sync simulation settings functionality in the context menu of the simulation subprojects.
Simulation results and analysis
The simulation results are automatically accessible in trenoLive for further analysis. Please go to the Analysis page to learn more.