This repository contains working code for the development of CMAP’s commercial services vehicle model (CSVM). An advanced commerical vehicle touring model using functions from rFreight to simulate commercial vehicle touring in the CMAP modeling region of NE Illinois and adjacent portions of Indiana and Wisconsin.
Minimum OS: 64-bit Windows 7, 64-bit Windows 8 (8.1) or 64-bit Windows 10 Processor: Intel CPU Core i5-2500K 3.3GHz / AMD CPU Phenom II X4 940 (or equivalent) Memory: 32 GB RAM
Recommended OS: 64-bit Windows 7, 64-bit Windows 8 (8.1) or 64-bit Windows 10 Processor: Intel CPU Core i7 3770 3.4 GHz / AMD CPU AMD FX-8350 4 GHz (or equivalent) Memory: 64 GB RAM
The CSVM runs in R 4.1.2, available from the CRAN R Project website at https://cran.r-project.org/bin/windows/base/R-4.1.2-win.exe.
The current release zip file of the CSVM, cmap_csvm-v1.0.0.zip is
installed and set up to run using the following steps:
cmap_csvm-v1.0.0.zip file from https://github.com/CMAP-REPOS/cmap_csvm/archive/refs/tags/v1.0.0.zipcmap_csvm-v1.0.0.zip into the parent directory that you wish to use for running the CSVM.htruck_congested_skim.omx, ltruck_congested_skim.omx, mtruck_congested_skim.omx) to the base scenario's inputs folder (scenarios/base/inputs)R that has been tested with this model, which is currently R 4.1.2, from the CRAN R Project website at https://cran.r-project.org/bin/windows/base/R-4.1.2-win.exe.lib/scripts/_USER_VARIABLES.R constains a setting USER_PROCESSOR_CORES. If this is set to a value more than 1, several of the steps in the model will run in parallel. While this reduces run time, it requires more memory, and, in certain circumstances, the overhead of parallel process data copying and management can reduce or remove the benefit of parallelization.
USER_PROCESSOR_CORES to 1. USER_PROCESSOR_CORES to 4. (More than 4 can be used although it has not been found to provide significant reductions in run time). run_cmap_csvm.bat will run the model when the batch file is double clicked or executed from the command line. The contents of the batch file are shown below. The command line arguments that are passed in the batch file identify the name and year of the scenario to be run, the reference scenario for comparisons in the dashboard, and the model steps to run. The scenario name must match the name of a scenario in the scenarios directory. A full list of command line arguments that can be passed to the model is described in run_cmap_csvm.R.:: CMAP Commercial Services Vehicle Model Batch File
:: Variable definitions
:: Scenario and Year
set scenarioname="base"
set scenarioyear=2019
:: Reference scenario and year to use for scenario comparison
:: Note: this scenario must have already been run
set reference="base"
set referenceyear=2019
:: Steps of the model to run (TRUE or FALSE, use upper case)
set runfirmsyn="TRUE"
set runcvtm="TRUE"
set runttexp="TRUE"
set rundashboard="TRUE"
:: For reference scenario, if this is the base scenario change to validation
if [%scenarioname%]==["base"] set reference="Validation"
if [%scenarioname%]==["base"] set referenceyear=2019
:: Run CSVM Model For Selected Scenario and Components
Rscript run_cmap_csvm.R %scenarioname% %scenarioyear% %runfirmsyn% %runcvtm% %runttexp% %rundashboard% %reference% %referenceyear% >run_cmap_csvm_log.txt 2>&1
:: Check for errors, exit and return error code if error
if %errorlevel% neq 0 exit /B %errorlevel%
:: Add pause to keep command window open at end of run
pause
The base scenario for the CSVM must be run before any alternative scenario can be run as the firm synthesis outputs from the base are required by alternative scenarios. The CSVM model is currently set up with the base scenario folder called "base". The selection of which scenario is the base scenario is a parameter setting, BASE_SCENARIO_BASE_NAME in lib/scripts/_BASE_VARIABLES.R.
To run only parts of the model for a scenario, some of the switches in the batchfile (runfirmsyn, runcvtm,runttexp, rundashboard) can be set to "FALSE" to not run that step. This is particularly helpful if you want to say run the dashboard again with a different reference scenario, when just the dashboard step can be run after updating the name of the reference scenario. It is impotant to be careful with these settings so as to not accidently skip a model step when it is intended to be run.
This section of the README discusses the scenario specific inputs to the model and how to adjust them to test out alternative scenarios.
Each scenario's inputs and outputs are contained in a directory called scenarios\[scenario name]. The set of input files required in each scenario are as follows:
cmap_csvm\scenarios\[scenario name]
* Inputs Scenario inputs directory
+ data_emp_control_taz.csv Land use data describing TAZ employment
+ data_hh.csv Land use data describing TAZ households
+ htruck_congested_skims.omx Heavy truck congested travel time and distance skims
+ ltruck_congested_skims.omx Light truck congested travel time and distance skims
+ mtruck_congested_skims.omx Medium truck congested travel time and distance skims
+ scenario_adjustments.R R script containing scenario specific parameter adjustments
* Ouputs Scenario outputs directory
A new scenario can be added to the model by copying (in File Explorer) a similar scenario that is most like the alternative scenario that is to be tested. For example, if you want to test a future scenario but with an alternative land use forecast, copy the scenarios\future directory and paste it into the scenarios directory. Then rename the folder to something like future_alt_landuse. Once the new folder is there the input files in scenarios\future_alt_landuse\inputs can be adjusted to reflect the specification of the scenario.
In order to test alternative land use scenarios, the files data_emp_control_taz.csv and data_hh.csv can be edited to increase or decrease the amount of employment of the number of households in each TAZ, respectively.
The model is sensitive to travel times based on the zone to zone travel times included in the skim files. In order to evaluate the impact of transportion projects or to test out a more general increase or decrease in region wide congestion, the skims files (ltruck_congested_skims.omx, etc.) should be updated (e.g., based on network updates and a new run of the travel demand model)
The script file scenario_adjustments.R allows the model user to assert adjustments to the parameters of the scheduled stop generation model and the vehicle choice model to evaluate higher and lower stop generation by activity, increases or decreases in the area serviced by a particular businesses, changes in sensitivity to cost of travel, and changing patterns in vehicle usage. The script is shown below. As noted in the script comments, positive values for the parameter adjustments will results in an increase in the measure (e.g., number of good stops, use of light vehicles)
In all cases, the adjustments are to the values of model parameters (e.g., constants and variables in logit models) and as such are not targets. Instead, small changes can be tested experimentally (for example, by making changes to the adjustments in increments of 0.1) to understand the downstream effects of a given changes in the number of stops, the distance to stops, or percentage change in vehicles choices.
# CMAP CSVM
# Scenario Specific Parameter Adjustments
# This script is read at run time and asserts adjustments to choice models in the CSVM structure
# Stops Generation
# Good deliveries (poitive adjustment = additional goods deliveries)
asc_goods_adj = 0.0
# Stops Generation
# Service calls (positive adjustment = additional goods deliveries)
asc_service_adj = 0.0
# Proximity of stops to businesses (to represent for example localization)
# Proportional adjustment (positive adjustment is a factor > 1, negative adjustment is a factor < 1)
# Since businesses (and all industries) do a mix of both activities, in application the factors are blended
# The effect of the goods factor on industries that have a higher emphasis on goods delivery is more pronounced and vice versea
base_dist_goods_factor = 1.0
base_dist_service_factor = 1.0
# Travel cost sensitivity adjustment
# This factors distance or time variables
# (representing generalized impendance in the stop generation model,
# a factor greater than one will reduce base to stop distance and also reduce overall stop generation,
# a factor less than one will increase base to stop distance and also increase overall stop generation)
impedance_goods_factor = 1.0
impedance_service_factor = 1.0
# Vehicle Type
# Mode specific constant adjustments by vehicle type (positive adjustment = trips shift towards that vehicle type)
asc_vehicle_light_adj = 0.0
asc_vehicle_medium_adj = 0.0
asc_vehicle_heavy_adj = 0.0
An alternative scenario is run in the same way as the existing scenarios provided with the model are run, using the batch file in the root of the CVSM folder, run_cmap_csvm.bat. The scenario name in the batch file must match the name of the new scenario created in the scenarios directory.
When an alternative scenario is run, you might want to compare the results to something other than the base scenario. For example, an alternative future land use scenario might be compared with the standard future scenario. To build the dashboard using the future scenario as the reference, change the reference scenario and reference year variables in the run_cmap_csvm.bat batch file as shown below:
:: Reference scenario and year to use for scenario comparison
:: Note: this scenario must have already been run
set reference="future"
set referenceyear=2050