Project Meeting 2020.12.17

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

• Meet next Tuesday the 22nd but cancel the 24th and 31st
• Still working on getting the multizone work released by end of year
• BYU meeting went well
• They also expressed interest in a simpler / more generic example model for typical MPOs
• Using the new dynamic chunking feature should help their runtimes
• Oregon DOT is a promising option to contribute the simpler / more generic example model to the consortium

5b Task 2 Strategic Planning (milestone)

• Partners discussed and agreed its good for me to now review/address comments
• The document will be a living doc, but we'll now wrap up this version as part of the contract deliverable
• We need a better articulation of the "spines/products/templates" and the versions/releases, requirements for setup, etc.
• And how this relates to help for developers as well as users/deployment
• Which spine should I pick?
• Let's better formalize the "spines" throughout the entire system

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

• Debugged and corrected a few more expressions/configs related to drive transit so results looking better
• Having the real full scale example is so helpful for verification of the code / example
• There's certainly a balance between perfecting this example we're using for testing versus project resources
• The code is working as expected
• Would be good to version the validation results so its easier to see the changes with each iteration
• Being able to debug/run just tour mode choice in 8 minutes and reload the cached TVPB utilities is super helpful

6a Task 2 Complete Estimation Mode (milestone)

• stop_frequency updated and now working on trip_destination and trip_scheduling

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

• No update

6a Task 8 Maintenance and Support (milestone)

• Providing a bit of support to Stefan as he works on getting the model running at PSRC
• Working on some memory issues for his parcel/maz setup
• We don't have a complete 40k maz TM2 activitysim test setup, but we are setting up the SANDAG cross border model, which uses 3 zones and TVPB and may require some improvements, which we will of course contribute
• Stefan to keep working on it and RSG to support
• Clint's efforts would be easier if we did a better job with backwards compatibility
• We already updated the TM1 example to use the updated code that supports multi zone systems
• Maybe we should think about versioning the expressions API?
• Supporting multiple zone systems and estimation integration resulted in a lot of core code changes
• Hopefully the code base will be more stable moving forward - i.e. post our push to version 1
• Let's be more intentional about backwards compatibility
• We'll hear more from Clint on Tuesday