Home

Bike Model

This is the draft WFRC bike model design.

Incremental logit bike demand model

• Runs at the zone-to-zone level

build_bike_trips = total_trips * [
  (base_bike_trips * exp(build_utility - base_utility) / 
  (base_other_trips + base_bike_trips * exp(build_utility - base_utility) ]

AMBAG incremental logit bike demand model

• Runs at the zone-to-zone level
• Input variables:
  • walk distance
  • bike generalized cost route utility
    • distance
    • distance on bike lanes
    • distance on bike routes
    • distance on arterials without bike lanes
    • distance on bike paths
    • number of turns
    • total gain in elevation
  • walk, bike intrazonal trip flag
  • motorized route utility (travel time)
  • alternative specific constants by trip purpose and income

Work flow

Base scenario steps:

• ArcGIS export multimodal network to CSV files with headers
  • nodes, links, link shapes, transit stops
• ArcGIS export microzone system to CSV files with headers
  • zones, zone shapes, land use
• Python script reads CSV files into pandas DataFrames and store in one HDF5 data store binary file for efficient I/O
• Cube export WFRC travel model zonal data to CSV files with headers
  • zones, zone shapes, land use data
• Cube export WFRC mode choice matrices and network LOS skims to OMX files
  • trips by mode by time period
  • auto travel time matrix by time period
• Python script adds z coordinate to every x,y coordinate in the HDF5 data store
• Python script disaggregates WFRC model zone-to-zone data to microzone-to-microzone
  • using microzone land use
  • including accessibility measures from multimodal network
• Python script generates walk and bike route costs using multimodal network

The result is microzone-to-microzone trips and route costs by mode for the base scenario.

Build scenario steps:

• User edits the multimodal network in ArcGIS
• ArcGIS export multimodal network to CSV files with headers
  • nodes, links, link shapes, transit stops
• ArcGIS export microzone system to CSV files with headers
  • zones, zone shapes, land use
• Python script reads CSV files into pandas DataFrames and store in one HDF5 data store
• Python script adds z coordinate to every x,y coordinate in the HDF5 data store
• Python script updates the disaggregation of WFRC model zone-to-zone data to microzone-to-microzone
  • using microzone land use
  • including accessibility measures from multimodal network
• Python script defines the area of influence for the build scenario in order to only run the route costs for a subarea to save time
• Python script generates walk and bike route costs using multimodal network
• Python script runs incremental logit model using the trips by mode and change in bike route utility
• Python script summarizes results and returns report to ArcGIS user

The result is microzone-to-microzone trips and route costs by mode for the build scenario.

Notes

• This PhD work will inform our design of the bike route choice model