Project Meeting 2020.11.12 - ActivitySim/activitysim GitHub Wiki

5b/6a Task 1 Project Management (milestone)

  • Oregon DOT is planning to multiprocess populationsim and add seed sharing for their statewide model and contribute the improvement
    • Stefan reminded the team to say multiprocessing instead of multithreading and shared a link
    • The statewide model runs multiple years into the future and runs for 50 to 60 PUMAs
    • Plan to multiprocess by PUMA
    • MetCouncil used seed sharing to add additional nearby PUMS samples to rural areas with not enough records in small markets
    • How do we decide how to share seeds - use neighboring PUMAs?
    • What's the process for rogue contributions like this? How do they align with our roadmap and the managing contributions section of the Strategic Plan?
    • These contributions sound good and are from trusted sources
    • Folks interested in the project should reach out to Alex
    • Let's use this as a test case for a more formal review process
  • Ohio DOT is in the process of joining!
  • As the user community grows, it might be good to do a semi-annual update for the broader community
  • And it would be good to align that with versions

5b Task 2 Strategic Planning (milestone)

  • Joe working on revisions in a new version and will share with the team soon

5b Task 10 Support for Three Zone Systems and TVPB (milestone)

  • For the full scale Marin TM2 example, I corrected the drive transit issue and have traced a couple ODs to check results
  • We added ability to re-trace tap-tap calcs when pre-computed since otherwise they do not get traced
  • I mapped the selected TAPs for the trace and they look good; also the trace calculations look good
  • I'm working through the set of review summaries we discussed
  • TM2 TVPB coefficients were estimated by assigning the onboard survey and comparing paths
  • Can we add a total utility column to the tracing? Maybe we add multiple levels of tracing for general and then detailed analysis. Should all time periods be in the tap-to-tap trace or just the relevant one?
  • Maybe we exponentiate ahead of time to save runtime
  • Jeff added a new submodel step - initialize_los - that pre-computes the tap-to-tap data
  • He's working on the shared memory data structures/access revisions for multiprocessing now
  • Inside mode choice we'll just be ranking best paths - not calculating all the detailed path utilities again and again
  • We can discuss the DaySim PNR lot choice with ActivitySim page I drafted next time

6a Task 2 Complete Estimation Mode (milestone)

  • Newman drafted the non-mandatory tour frequency interaction_simulate larch notebook
  • Newman working on wrapping it with a run_all_steps feature and also revising how the 8 EDBs by person type should be merged into 1 EDB like the other submodels

6a Task 3 Telecommute Model in Cooperation with SEMCOG (milestone)

  • Telecommute design and links to background docs posted
  • Generally looks good
  • Can we make it any easier for the user to do what-if analysis? Maybe some kind of smart constant calculator for target telecommute shares or something like that?
  • I plan to draft spec/example for Blake to start programming
  • Any additional comments due next week

6a Task 8 Maintenance and Support (milestone)

  • Joe talked with Beam/LBL, BYU, and hasn't heard back yet from Toledo
  • Caliper is helping Toledo with their setup
  • Joe plans to set up a call with these users to learn more about their experience
  • Need to review Clint's ARC CDAP and Duplicate Time Labels PR
  • Need to review Albab's SEMCOG joint_tour_destination configs PR