Scribd
Upload
34 views

Example of Stratified Diversion-Curve Model

1) The document provides an example of applying a stratified diversion curve model to estimate transit ridership. It estimates that 2000 peak hour trips will be made, with 200 by captive riders, 600 by income group EC3, and 1200 by income group EC5. 2) It then calculates the transit share using the diversion curve, finding 590 person-trips would be made by transit. 3) It also discusses discrete choice models like multinomial logit that are used to model travel behavior, describing aspects like the utility function and how probabilities are calculated.

Uploaded by

shoumik
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
34 views

Example of Stratified Diversion-Curve Model

1) The document provides an example of applying a stratified diversion curve model to estimate transit ridership. It estimates that 2000 peak hour trips will be made, with 200 by captive riders, 600 by income group EC3, and 1200 by income group EC5. 2) It then calculates the transit share using the diversion curve, finding 590 person-trips would be made by transit. 3) It also discusses discrete choice models like multinomial logit that are used to model travel behavior, describing aspects like the utility function and how probabilities are calculated.

Uploaded by

shoumik
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Example of stratified diversion-

curve model
 A peak-hour target year trip interchange volume between two zones has
been estimated to be 2000 person-trips as follows: 200 of these trips are
made by captive riders, 600 of these are made by trip makers from the
household of income category EC3 and the remaining 1200 from the
income category EC5. Apply the Washington D.C. stratified model to find
the market share of the transit system assuming that this model has been
calibrated as a disaggregate model based on household income
categories. The following data apply to the interchange under
consideration: TTR=2.0, CR=0.8 and L=4.0
Figure of Stratified-Diversion Curve
Figure of Stratified-Diversion Curve (Contd.)
Solution
• Transit-captives: 200 person-trips
• EC3 group: 600 × 0.25=150
• EC5 group: 1200× 0.20=240
• Total transit share=200+150+240=590 person-trips
Disaggregate Method (Discrete choice model)

Widely used discrete choice models are


• Multinomial logit model
• Nested logit model
• Multinomial probit model
Multinomial Logit Model
• The utility or disutility function is typically expressed as the linear sum
(wighted sum) of the independent variables
U=a0+a1x1+a2x2+……………+anxn
• U=utility
• a0=alternate-specific constant/ mode-specific constant
• a1, a2 ………….an=coefficients
• x1, x2, ………., xn=variable representing the attributes of the choice or of
the traveller
Approaches for formulating the utility equation for Multinominal Logit
Model

Two approaches
• Mode-specific/ choice-specific approach
• Attribute-specific / choice-abstract approach
Mode-specific and Attribute-specific formulation of the utility equation

• Mode-specific/ choice-specific approach


• U1=6.2+2.4x1+3.5x2
• U2=3.4+3.1x3+2.9x4
• U3=4.3+2.9x5+3.2x6
• Attribute-specific / choice-abstract approach
• U1=6.2+2.4x1+3.5x2
• U2=6.5+2.4x1+3.5x2
• U3=7.0+2.4x1+3.5x2
Multinominal logit model
Probability of choosing a particular mode

P(i)=eu(i)/∑eu(i)
Example of multinominal logit model

A calibration study resulted in the following utility equation


Uk=ak-.025x1-.032x2-.015x3-.002x4
where
• x1=access time in minutes
• x2=waiting time in minutes
• x3=line haul time in minutes
• x4=out-pocket cost in cents
Example of Multinominal Logit Model (contd.)

• The trip forecast for a particular interchange was a target-year volume of Qij=5,000 peson-
trips per day. During the target year, the trip-makers on this particular interchange will have a
choice between the private automobile (A) and a local bus system (B). The target year service
attributes for the two competing modes have been estimated to be
Attributes x1 x2 x3 x4
A 5 0 20 10
B 10 15 40 50
• Mode specific constants are aA=-0.12 and aB=-0.56. Estimate the market share of the two
modes and the resulting revenue of the bus system
Introduction of A new mode
Find the utility of each mode
• U(A)
• U(B)
• U(C)
Find the probability of choosing each mode
• PA=eU(A)/(eU(A)+eU(B) +eU(C))
• PB=eU(B)/(eU(A)+eU(B) +eU(C))
• PC=eU(C)/(eU(A)+eU(B) +eU(C))
Sensitivity to other policies
Probability of choosing a mode under existing conditions
• U(A)
• PA=eU(A)/(eU(A)+eU(B) +eU(C))
• Probability of choosing a mode under changed scenario
• U/(A)
/
• P A=eU/(A)/(eU/(A)+eU(B) +eU(C))

You might also like