Skip to content

Run the Model

Cundo Arellano edited this page Apr 15, 2020 · 13 revisions

This section provides instructions for executing the travel model. For a description of the underlying computing environment, see Computing Environment; for a general description of the underlying system design, see System Design; for a description of the configuration files likely needing to be modified before executing the model, see Setup and Configuration.

Execute Entire Model Stream via Emme

One can run though the entire model stream from the CT-RAMP java-based models through assignment from within Emme. This starts the household manager, matrix manager, JPPF driver, builds skims, runs all model components, and performs highway and transit assignments. The current SANDAG implementation runs three full iterations of the models (i.e. the model starts with assignment of base-year trip tables, and speeds from previous assignments are fed back through the entire model stream).

The SANDAG travel demand model is executed via the Master run tool, which can be run from Emme Modeller. To access the Master run tool, start from a project configured as described under Setup and Configuration.

  1. Open Emme Desktop using the emme_project/emme_project.emp file. Double click on the file, see figure below.

  1. Highlight Scen. 100 (-- --): Empty Scenario and click OK.
  2. Open the Emme Modeller window, from the icon in the toolbar, or from the menu under Tools -> Modeller. If the icon is greyed out, it’s likely due to unsuccessful creation of emmp_project due to invalid Emme license.
  3. From the Modeller window, open the SANDAG toolbox in the left-hand pane, and double-click on the Master run tool as shown in the figure below.

  1. The Master run tool operates the SANDAG travel demand model. To operate the model, configure the inputs by providing Scenario ID, Scenario title and Emmebank title, and keeping Number of Processors to default. Select main ABM directory will automatically be set to the current project directory and does not require change. The Master run tool UI is shown in in the figure below.

    • Main ABM directory: directory which contains the ABM scenario data, including this project. The default is the parent directory of the current Emme project.
    • Scenario ID: Scenario ID for the base imported network data. The result scenarios are indexed in the next five scenarios by time period.
    • Scenario title: title to use for the scenario.
    • Emmebank title: title to use for the Emmebank (Emme database)
    • Number of processors: the number of processors to use for traffic and transit assignments and skims, aggregate demand models (where required) and other parallelized procedures in Emme. Default is Max available – 1.
    • Properties (conf/sandag_abm.properties)
      • On opening the Master run tool the sandag_abm.properties file is read and the values cached and the inputs below are pre-set. When the Run button is clicked this file is written out with the values specified. Any manual changes to the file in-between opening the tool and clicking the Run button are overwritten.
      • Sample rate by iteration: three comma-separated values for the CTRAMP sample rates for each iteration
      • Start from iteration: start iteration for the model run
      • Skip steps: optional checkboxes to skip model steps.
    • Select link: add select link analyses for traffic.

  1. Expand the Run model – skip steps drop down to make any custom changes. Usually the defaults should be sufficient although if you are using a new bike network, you should uncheck the Skip bike logsums and check the Skip copy of bikelogsum

  1. Click Run to start the model run.
  2. Occasionally check model run status. If successful, the model will finish in 2.5 to 3 days.
  3. Close out of Emme windows and log off server if finished.

If the run is unsuccessful (there will be an error prompt from Emme), check Emme logbook and log files (under “logfiles”) for clues to where it stopped.

As the model runs, a full runtime trace of the model steps, inputs and reports is recorded in the Modeller Logbook. As each step completes, it will record an entry in the Logbook along with reports and other results. The Logbook can be opened from the icon in the upper right of the Modeller window. This icon can also be found in the toolbar in the Emme Desktop. If a Modeller tool is running, a window will pop-up over the Emme Desktop which includes a Show Logbook button (this window can be closed to use Desktop worksheets and tables while any tool is running). Click on the Refresh button to update the logbook view with the latest status.