Project Meeting 2025.06.24

Agenda

• Explicit Error Terms

Technical

• Explict Error Terms
  • Software enhancements complete
  • Runtime analysis
    • Runtime with explicit error terms is 2.4x the runtime of MC simulation in full SANDAG model
    • Vast majority of additional runtime is in destination choice model (96% of increase)
  • Tests conducted
    • Land-use scenario test
    • Transit scenario test
  • Does EET simulation result in more logical choice outcomes at a disaggregate level than MC simulation?
    • Auto ownership: transit scenario expectations
      • See more households reducing auto ownership with MC simulation than EET simulation.
      • Change in auto ownership with MC simulation is always +/- 1 car, whereas in EET there are changes between nonadjacent car choices.
      • Neither EET nor MC had any changes in auto ownership for HHs outside the affected area
      • Overall, very small changes
    • Mandatory location choice: transit scenario expectations
      • Affected area, 0.6% of work locations change in EET, vs 3% in MC
      • Unaffected area, no work location changes
      • Affected area, 3.2% of school location changes in EET, vs 3.1% in MC
      • Unaffected area, 6.5% of school locations change in EET, vs. 7.6% in MC
      • Overall the changes in mandatory location choice suggest more stability in the EET model w.r.t. changes in the affected area for work, but relatively similar changes for school location across EET and MC
  • Tour mode choice: transit scenario
    • Changes in mode is more reasonable for EET
  • Visualization tool has many summaries
    • Tool will be distributed once formatting is fixed
    • Tool takes half hour for each scenario to load after each change
    • Maps would be nice, but not critical