Transport Demand - Trip Generation Flashcards
What is the 4 - Step Transport Model?
Inputs:
Land Use Data
Travel Data
Zones and Networks
---------
Process:
--->Trip Generation
--> Trip Distribution
--> Modal Splits
--> Assignment
---->Evaluation
What is a trip?
One way movement from A to B
What is a Home based trip?
A trip with one end at home
What is a non-home based trip?
A trip with neither end at home
What is trip purpose?
Determined by activity at the non-home end of a home based trip
What percentage of trips start or end at home?
85%
What is trip production?
The home end of the HB trip or the origin of a NHB trip
What is trip attraction?
Non home end of a HB trip or the destination of a NHB trip
What is trip generation?
The process of estimating productions and attractions for each zone in the model
What factors affect trip generation?
Trip production:
- household size
- income and car ownership
Trip attractions:
- population
- employment
- enrolements
What are the three types of model forms?
- Growth Factor Modelling
- Category Analysis
- Regression Analysis
What is growth factor modelling?
Ti = Fi ti Ti= future trips to/from zone i ti = existing trips to/from zone i Fi = growth factor (depends on population, income, car ownership)
Relies on a pre-existing estimate of productions and some measure of growth
What is category analysis?
Estimates average trip rates as a function of household attributes (common on UK and US)
Categories should:
- be large enough for sufficient data
- minimise intra category variation in trip rates
Issues:
- no statistical goodness of fit
- cannot extrapolate
- no effective way of choosing variables
What is regression analysis?
A dependent variable is estimated from a set of independent variables
Yi = ∑ bk Xki + c + ei
Where bk = the coefficient for the kth parameter
Xki = the kth independent variable for zone i
c= constant or intercept
ei = the error or disturbance term
What are the three stats test that are important?
- t-test = significance of independent variables
- F -test - significance of the whole equation
- R^2 = proportion of variation explained
What is zone based regression?
An equation that deals with zonal total trip ends - based on zonal attributes
- have a large constant term
- coefficients depend upon the zone system used
- often used for trip attraction modelling
e.g.
Trips = 1.23 X populationi + 1.10xEmploymenti + 750
What is household based regression?
Deal with household trips ends
Trips = NumHHi ( … x Persons/HH + .. Workers /HH+ … Cars/HH + …)
Fewer independent variables gives better relationship forms
Relationships tend to be geographically and temporally stable
Greater variation between households
What is dummy variable regression?
Dummy variables - take on a value of 0 or 1, depending on the absence or presences of pre-defined effect
- Capture non-linearity between trip rates and independent variables
- Developed at the household level but applied at the zonal level
Must Productions and attractions equal? and if so why is this required?
If attractions dont equal productions what is the pragmatic solution?
Yes, there is no guarantee that they will but the ∑Ai = ∑Pi
Required because conservation of matter and for trip distribution
Pragmatic solution is to factor Attractions (Ai) by a factor f to equal productions
What are the implementation issues for models?
- Assume that model parameters do not change with time ( once calibrated model coefficients remain constant)
- Assume geographically stable
- Challenges remains to forecast the various input variables
- Trip generation modes ignore the impact of transport supply, congestion and accessibility on trip making
What are the stability parameters?
- Temporal Stability
- Geographic stability
How are input variables forecasted?
- Production variables (number of households, average people, workers, dependents and cars per household)
- Attraction variables (employment, enrolments)
True or False
Any errors in forecasting input variables will flow through the rest of the 4-step process
True
Why is trip generation considered inelastic?
- the network has no effect on trip generation
- unlikely to be true for discretionary trips
- one possible solution
- incorporate accessibility in trip generation model - but research has found that accessibility is not significant in models
