Project Meeting 2019.05.17 - ActivitySim/activitysim GitHub Wiki

Verification

  • See the verification table for the current summaries, which keep looking better and better
  • Issues we fixed since our last call, which are in pull request PR#261
    • Add at least one non-preschooler to filter for HHs that participate in the joint tour freq
    • Calculate time overlaps for all persons in HH for joint tour freq
    • Correct ptype, cdap expressions for calculating HH time overlaps
    • Correct use of emp status (full or part time) instead of person type in atwork tour freq
    • Correct non-worker indicators in atwork tour freq
    • Add constants for preschoolers for non-man tour freq
    • Correct expressions offset error for school and preschool ptypes for non-man tour freq
    • Add constants for joint tours for non-man tour freq
    • Turn on stop freq num stops by purpose expressions
    • Correct income expression for stop freq atwork
    • Correct max distance by purpose for stop freq
    • Correct tour mode choice IVT coefficients by transit submode
    • Correct auto ownership filter for transit for tour mode choice
    • Drive alone not available for escort for tour mode choice
    • Fix transit origin and dest walk times for tour mode choice
    • Use offpeak parking costs for non-man for tour mode choice
    • Divide transit skims by 100 (because they are in Cube format) in tour mode choice
    • Correct trip mode choice IVT coefficients by transit submode
    • Set trip mode choice max walk to 1 mi instead of 1.5 like tour mode choice
    • Fix transit origin and dest walk times for trip mode choice
  • Re-ran TM1 with comparable setup, which includes these additional UEC fixes, see PR#22
    • IndividualMandatoryTourFrequency.xls - not divide auto skims by 100
    • JointTours.xls - moreCarsThanWorkers calculated incorrectly
    • ModeChoice.xls - Drive transit coefficients on IVT incorrect
    • TripModeChoice.xls - correct walk mode walk time max
  • Still working on reviewing tour stop frequency
  • Plan to publish (release) an updated version after that
  • Likely do one more pass through to correct additional differences
  • Updated verification script to run for either ASIM (or ASIM and TM1)

Admin

  • Welcome Dennis Farmer from MetCouncil
  • @Joe follow-up with agency partners about implementation plans for workshop; @Guy share ideas
  • @Jeff working on updating our multiprocessing runtimes for our TRB Planning Applications conference paper
  • @Ben to draft SOW for next phase of work. I'm planning to scope the following tasks for consideration:
    • Features
      • Support for multiple zones
      • Support for AVs/TNCs
      • Incorporation of PopulationSim into the consortium
      • Parking location choice
      • OD demand matrix builder
    • Usability
      • Improved input validation / checking for easier use
      • Output reporting / summarization functionality for easier use
      • Additional user documentation / tutorial with exercises in the wiki
    • Additional performance tuning (i.e. optimization)
    • DaySim like estimation mode to output alternatives, logsums, etc
  • Project wiki
    • @Ben to add the Phase 1 (Synthicity) materials to the wiki
    • @Ben to restructure wiki for new users - move current stuff up, move historic stuff down
    • @Ben add a possible future features page with ideas, ranking, and target release date for @Alex and others to contribute