360 likes | 437 Views
Examination of a Quality Control Forecast Model for Transit New Starts Projects. Arash Mirzaei P.E. Huimin Zhao Ph.D., P.E. North Central Texas Council of Governments. 11 th TRB National Transportation Planning Application Conference. May 9 th , 2007. Acknowledgement. Ken Cervenka, NCTCOG
E N D
Examination of a Quality Control Forecast Model for Transit New Starts Projects Arash Mirzaei P.E. Huimin Zhao Ph.D., P.E. North Central Texas Council of Governments 11th TRB National Transportation Planning Application Conference May 9th, 2007
Acknowledgement • Ken Cervenka, NCTCOG • Jim Ryan, FTA • Nazrul Islam, FTA • DCTA Project Team
Disclaimer • Work in Progress • Preliminary Results and Conclusions
Outline • Objectives • Mode Choice Model at NCTCOG • SUMMIT for New Starts • Incremental Logit Model • Transit Skim • Comparison Results and Lessons Learned
Objectives • To be Compatible with FTA Guidance and Recommendations • To Test the Semi-Independent Ridership Forecasts (QC Model) for New Starts • To Compare the User Benefits between a Locally Developed Transit Skim/Mode Choice Model and a Quality Control Model
Mode Choice Model at NCTCOG • Model Structure • Nested Logit Models for HBW and HNW • A Multinomial Logit Model for NHB • Stratified Sample Data • 1996 Household Travel Survey • 1996 FWTA On-Board Survey • 1998 DART On-Board Survey
Validation: YR1999 1999 Rail Station Ons and Offs (Weekday)
Comparison of Daily TRE Ridership: 2005 Observed vs. NCTCOG Model 2005 NCTCOG Model Validation: YR2005
2005 TRE Rail Station Ons plus Offs (Weekday) Validation: YR2005 (cont)
Mode Choice Model (cont.) • Nesting Structure for HBW
NCTCOG Mode Choice Model Structure • Nesting Structure for HNW
P P1 ΔP P0 ΔX X X1 X0 Gradient ΔP/ ΔX Quality Control Model • FTA Recommends the Quality Control Alternative for a Commuter Rail New Starts Project in DFW Area • Incremental Logit Model
Impedance Weight for Transit Skim The Weights are Consistent with Mode Choice Model Coefficients
Start Base Case Transit System Alternative Case Transit System Transit Skim Using FTA Recommended Parameters NCTCOG Travel Demand Model Base Skim Matrix Alternative Skim Matrix FTA Coefficients Difference Matrix: Skim Variables, Utilities Base Case Matrix: Transit Trips, Utilities Alternative Case Matrix: Transit Trips, Utilities End QC Model Flow Chart
QC Model Procedure • Run NCTCOG Skim and Mode Choice on Base Case to Obtain Base Case Transit Trips and Utilities • Run Transit Skims on Base Case and Alternative Scenario Using FTA Recommended Weights • Calculate IVTT, OVTT, Fare, as well as Utility Differences between Base and Alternative Transit Skims • Calculate Transit Share Change due to Skim Differences and Obtain the Transit Trips for Alternative Case • Two QC Models are Tested: One with No Rail Constant, One with a 12-minute Rail Constant
Transit Skim Settings • Maximum Trip Cost • Maximum Number of Transfers • Maximum Walk/Drive Time • Maximum Initial/Transfer Wait Time • Combination Factor
User Benefit Comparison • Tested One Market Segment of HBW Trips: H23WVLTP • Compared Three Scenarios: NCTCOG Model, QC Model with No Rail Constant, QC Model with a 12-minute Rail Constant
Transit Drive Trip Differences: QC Model with No Rail Constant
Conclusions • QC Model Approach Offers a Level Playing Field to Compare New Start Projects across the Country • Transit Trips and User Benefit from QC Model is Much Less than Those from NCTCOG Model • Though Different in Quantity, We Observed Similar Patterns in Transit Trip Shifts, Especially for Transit Walk Trips, from Both Approaches
Next Steps • More Tests to Confirm the Implementation Procedure • Investigate Transit Skim Differences • Investigate the Impact of Differences in Transit Accessibility between Base and Alternative Cases