Project Meeting 2024.11.05

Agenda

• Technical - Estimation Enhancements Updates
• Admin - Phase 10A tasks

Technical

• Estimation Enhancements Updates
  • Two zone system has been tricky, so working through details to make sure code is working correctly
  • Larch update - running MTC one-zone data through Larch using parquet files

General Admin

• Phase 10A tasks
  • Tracing Memory Improvements with Sharrow #754
    • Need to get Sharrow data to write out to disk; Means of doing so is complicated and not straightforward
    • ARC marked it as a high priority, but OK with punting it; no immediate emergency
    • Can debug w/o tracing but have to get into debugger mode and stop execution in middle of running it - higher developer skillset to be able to do it
    • Level of effort is high, and uncertainty in achievability
    • 10A Decision: Punting
  • Improve Joint Tour Participation Model #569
    • Translink had issue during calibration; can not calibrate model directly and must run all three models jointly and use trial and error approach to hit number of joint tours that are reasonable by different segmentations
    • For other models you can calculate a target and adjust parameters directly until you hit the target; can use algorithm to get it to converge, but w/ joint tour model can not do that
    • Two task options
      • Quick fix to make it easier to calibrate
      • Design task to make joint tours better overall
    • 10A Decision: Punting
  • Input Checker Enhancements #894
    • Checker could catch most likely issues that occur
    • Duplicate of Item 9/#784 and Item 25/#873
    • Phase 10A: Combine all of these into single "project" and keep on short-list
  • Preprocessors and Annotators available for every model #895
    • Not available for location model; should be standardized across different models
    • Testing is what will take time
    • Phase 10A: Keep on short-list
  • Multi-year development roadmap
    • Deliverable would be integration of all materials that relate to consortium and roadmap; Update as necessary given all work that has been done since original roadmap
    • Phase 10A: Keep on short list
  • Very long time to write skims to shared data buffer when using sharrow
    • Solved a specific performance issue, but never got an answer to why it was bad in the first place
    • High level effort, uncertainty around solution; kind of have a work around
    • Ultimately should be fixed in open matrix
    • Phase 10A: Punting
  • Improve Data Model.. #728 and Level-of-service Manager #743
    • Level of effort high
    • Should be on roadmap, but may be too large for Phase 10A
    • Phase 10A: Punting
  • Testing Required for various directory solutions #902
    • Should be done, but can be rolled into maintenance and support
    • Phase 10A: rolled into maintenance and support; do not write up as separate task
  • SEMCOG Model taking longer to run using ActivitySim 1.3.1 with Sharrow... #904
    • High on usability improvement level of importance
    • Relates to Item #4/#906 - specification performance analysis tool
    • Medium level of effort
    • Phase 10A: Short-list
  • Got through list of everything David H ranked as medium
    • Those on David H's "Low" list can be addressed in the support task; should be prioritized so they can be addressed in order of importance, ideally
  • Need scope descriptions for items on short-list; Will reserve one meeting (Nov 12) for technical discussion on scope