Travel Demand Forecasting: Parameters and Techniques

37  

In model application, these probabilities would be com- always estimate zero trips when the values of the explanatory
puted and applied separately to segments of households of variables are all zero. Mathematically, this is equivalent to
each type as defined by the variables (number of persons, saying that the trip generation equations should include no
income level, etc.), and the probabilities for each segment constant terms.
applied to the households in each segment.
Because no two of the models presented in Tables C.2
4.4.1  Model Function
through C.4 have identical specifications, the values for spe-
cific coefficients may differ significantly between models. The The purpose of trip generation is to estimate the num-
presence or absence of other variables in a model can affect ber of average weekday trip ends by purpose for each zone.
the coefficients of other variables. So it is much more valid to In four-step models, the trip ends of home-based trips are
transfer individual models rather than composites of models defined as productions, representing the home ends of trips,
with different variables. and attractions, representing the nonhome end, regardless of
As discussed previously, there is little experience with whether home is the origin or destination. In other words, for
which to guide planners in transferring vehicle availability home-based trips, the production end may be the destination
models, or even to determine how transferable the param- and the attraction end, the origin if the trip-maker is return-
eters of such models are. The best guidance that can be pro- ing home. For nonhome-based trips, for convenience the
vided if one wished to transfer one of the models shown in production end is defined as the trip origin and the attraction
Tables C.2 through C.4 is to choose one of the models based end as the trip destination.
on the similarity to the metro areas based on the charac- For home-based trips, the number of trip productions in a
teristics provided above (location within the United States, zone is, naturally, based on the number of households in the
population, and average vehicles per household). Because of zone. Household characteristics can affect trip making; there-
the differences in model specification, a composite of two or fore, in trip production models, households are usually clas-
more of these models cannot be created. If the chosen model sified by some of these characteristics, which often include
proves difficult to calibrate, perhaps another model could be the number of persons, workers, children, or vehicles, or the
chosen for transfer. household income level. The trip rates for each purpose vary
depending on the household classifications, which may not
be the same for all trip purposes.
4.4  Trip Generation
Trip attractions are based on other variables besides
Trip generation is commonly considered as the first step households, because several types of activities (commercial,
in the four-step modeling process. It is intended to address employment, residential, etc.) are often located at the non-
the question of how many trips of each type begin or end in home trip end. The type of activity that affects the number
each location. It is standard practice to aggregate trips to a of trip attractions depends on the trip purpose. For example,
specific unit of geography (e.g., a traffic analysis zone).5 The home-based work trip attractions are usually estimated best
estimated numbers of trips will be in the unit of travel that by using employment as the explanatory variable. Other
is used by the model, which is usually one of the following: purposes typically use different sets of variables (school
enrollment or employment for home-based school trips,
• Vehicle trips; retail employment for home-based shopping trips, etc.).
• Person trips by motorized modes (auto and transit); or Home-based nonwork, home-based other, and nonhome-
• Person trips by all modes, including both motorized and based trip attraction models usually use a linear combina-
nonmotorized (walking, bicycling) modes. tion of several different variables (employment by type,
households, etc.).
Trip generation models require explanatory variables The number of nonhome-based trips made in a region
that are related to trip-making behavior and functions that does depend on the number of households, but unlike
estimate the number of trips based on these explanatory home-based trips, they need not have one end in the zone
variables. While these functions can be nonlinear, they are where the household of the trip-maker is located. One way in
usually assumed to be linear equations, and the coefficients which models deal with this issue is to use household-based
associated with these variables are commonly called trip nonhome-based trip production rates to estimate regional
rates. Whether the function is linear or nonlinear, it should productions and to allocate this regional total to zones based
on other variables. A common convention is to assume that
the regional nonhome-based trips are allocated to each zone
5While the geographic units of some travel models are not zones, the based on the number of nonhome-based trip attractions in
term “zones” is used in the remainder of the chapter for convenience. the zone.

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

38

Special Generators not be equal because productions and attractions are esti-
mated separately. While trip distribution models (see Sec-
While estimates of passenger trip activity based on rates
tion 4.5) can often be applied with different production and
applied to household or employment in a zone can address
attraction totals, certain types of model formulations (such as
the majority of conditions, there are special conditions
the gravity model) produce better results if productions and
when these rates are insufficient to accurately estimate trip
attractions are equal, or close to equal.
activity. These conditions might be because the trip activity
Because trip productions are estimated for the household,
is due to considerations not directly related to the number
which is the same as the basis of the sampling frame of the
of employees or households in a zone—for example, trips to
surveys from which trip generation models are estimated,
airports, hospitals, colleges, or large recreational facilities. Addi-
trip production models are generally estimated using records
tional estimates of trip activity may also be necessary because
representing individual households, for which the total num-
the trip generation rates are for average conditions that are
ber of trips should be reported in the household survey. Trip
not applicable to specialized conditions—for example, shop-
attractions, on the other hand, occur at locations for which
ping productions or attractions to “big box” retail stores that
a complete set of survey records comprising all trips to the
have shopping trip rates per employee that are higher than
attractor will not be available. It is therefore common con-
typical retail employment. These activity locations are often
vention to adjust trip attractions to match productions by
referred to as “special generators.”
purpose at the regional level. This “balancing” of productions
The term “special generators” is somewhat misleading in
and attractions must take into account trips with one end
that the different travel behavior associated with them is not
outside the region (see Section 4.6 on external travel) and
limited to trip generation. While it is true that the number
trips attracted to special generators.
of trips generated by these sites is not readily modeled using
It is good practice to review the ratio between unbalanced
conventional trip attraction models, the sensitivity of trip dis-
attractions and productions as a large difference might indi-
tribution (see Section 4.5) and mode choice (see Section 4.7)
cate problems with employment estimates, trip rates, etc.
to variables such as time and cost is also different than that of
Most literature on best practices recommends that the differ-
other trips. Ideally, such travel should be treated as a separate
ence between unbalanced regional attractions and produc-
trip purpose so that separate models for trip generation, trip
tions be kept to +/-10 percent for each purpose, although a
distribution, and mode choice could be applied, but unless
review of model validation reports shows that this standard
there are detailed surveys of the special generator with a suf-
is often exceeded. Upwards of +/-50 percent difference at the
ficient sample size for model estimation, it is unlikely that regional level might be considered acceptable under certain
this could be done. conditions and trip purposes.
Trip rates are not developed for special generators. Rather,
the numbers of trips attracted to these locations are exoge-
nously estimated using separate data sources, such as surveys 4.4.2  Best Practices
or counts conducted at the special generators. Hence there
Trip Productions
are no parameters for trip generation at special generators,
and default parameters cannot be provided. It is important to While other model forms are sometimes used, the most
consider how special generator travel is considered relative to common form of trip production model is the cross-classifi-
the trip purposes used in the model. Generally, trips attracted cation model. The households in each zone are classified by
to special generators are estimated separately from the attrac- two or more variables, and the number of households in each
tions for the trip purposes used in the model, but the special category is multiplied by the appropriate “trip rate,” repre-
generator attractions must be considered in examining the senting the average number of trips per household for the
balance between productions and attractions. Since separate category. Mathematically, the number of trips generated in
trip distribution, time-of-day, and mode choice models are a zone is given by:
not available for special generator travel, the analyst must
decide how these features will be modeled for special genera- P ip = ∑ P rate pk  hik (4-7)
tors (for example, using the models for home-based nonwork k

or nonhome-based travel).
where:
P ip = Number of trip ends produced for purpose p in
Balancing Productions and Attractions
zone i;
The regional totals of productions and attractions for each P ratepk = The production trip rate for purpose p per house-
trip purpose are equal because each trip has one production hold for category k; and
end and one attraction end. However, the model results may hik = The number of households in category k in zone i.

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

39  

The state of the practice for trip production models is to To summarize, the model parameters for trip generation
create tables of trip rates by two or more dimensions, for are the trip production and attraction rates, represented by
example by household income and by household size (num- Pratepk in Equation 4-7 and Aratepk in Equation 4-8.
ber of persons). Most commonly, trip production models are
two-dimensional, although three-dimensional models are
4.4.3  Basis for Data Development
sometimes used, especially in larger areas where more data are
available. The households in each zone are segmented along the When sufficient local data are available, best practice for
two dimensions, and the trip rate is estimated for each combi- the development of trip generation models is to estimate the
nation of the two variables. For example, a cross-classification model parameters from household activity/travel survey data
of households by three income levels (say, low, medium, and using statistical techniques such as linear regression. Typi-
high) and number of persons (1, 2, 3, and 4+) would have the cally, sample sizes for these surveys are sufficient for model
number of households divided into 12 segments, one for each estimation, although the required amount of data depends
income level–number of persons combination, and would use on factors such as:
12 corresponding trip production rates.
• The number of parameters to be estimated, such as the
number of cells in cross-classification models;
Trip Attractions • The number of households occurring in each cross-
Accurately estimating trip attractions can be significantly classification cell in the population, and in the survey
more difficult and problematic than estimating trip produc- sample; and
• The resolution of the geographic units (e.g., zones) at
tions. Whether trip attraction model parameters are estimated
from local data or are transferred, they are usually derived which the models will be applied.
from household survey data, which collects travel information
at the production end of trips. Such surveys do not provide If local data for model estimation are not available, param-
eters may be transferred from another model. Transferable
control totals at trip attraction locations. It is common practice
parameters for general use are presented in Section 4.4.4.
to estimate the parameters, such as coefficients in linear regres-
sion equations, at an aggregate level such as districts (groups
of zones), implying that the results may not be as accurate at Trip Productions
more disaggregate spatial levels (such as zones). Some regions
For trip productions, cross-classification trip rates were
have attempted to address this issue through the use of estab-
estimated from the 2009 NHTS for the classic three trip pur-
lishment surveys, where the data are collected at the attraction
poses, for urban areas stratified by population. Additionally,
end of trips, but the wide variety of establishment types and the
trip rates for home-based school trips are presented, along
expense of obtaining sufficient sample sizes at each type means
with a home-based other trip purpose that represents all
that accuracy issues are not completely resolved. It is therefore
home-based nonwork and nonschool trips. These rates rep-
recommended that analysts use the information provided here
resent average weekday person trips, including both motor-
(indeed, locally derived trip attraction information as well)
ized and nonmotorized trips, and were estimated using the
with extreme caution and to be prepared to adjust parameters weighted NHTS data. Initially, separate rates were estimated
to produce more reasonable results as needed. for the six urban area population ranges, but, in many cases,
Trip attraction models are most often linear equations the rates did not vary by population category, and combined
with variables representing the amount of activity in a zone— rates for multiple population ranges are presented.
typically employment by type, student enrollment at school Note that the 2009 NHTS does not include travel for
sites, and households or population—and coefficients reflect- children younger than five years old. If an analyst wishes to
ing the effects of these variables on trip making to the zone model the travel of younger children and to use the informa-
for the appropriate purpose. The equations follow the form: tion provided in this chapter, he/she should be prepared to
slightly adjust the trip rates for all purposes except home-
Aip = ∑ A rate pk  v ik (4-8) based work upward, with a more substantial increase in
k
home-based school trips (if that purpose is modeled and
where: includes pre-school/day care travel).

Aip = Attraction of trip ends for purpose p in zone i;

Trip Attractions
Aratepk = Rate of attraction trip ends for purpose p per unit
of variable k; and Documented trip attraction models from a number of
vik = Value of variable k in zone i. MPOs were available in the MPO Documentation Database.

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

40

One conclusion from the review is that there is little com- • $25,000 to $49,999;
monality among MPOs regarding the variables to include in • $50,000 to $100,000; and
trip attraction models. The variables ranged from employ- • Over $100,000.
ment stratified by three basic groups to employment strati-
fied by seven or eight groups. In a number of trip attraction To determine which variables best explained trip gen-
models, school enrollment was included. The number of trip eration behavior in the NHTS data, an analysis of variance
purposes and the variables used for each trip purpose also (ANOVA) was performed to explore the explanatory power
varied substantially. of the variables. This parametric statistical technique pro-
Different model calibration methods also added to the vari- vides a basis to identify the most statistically significant cross-
ation among models. Some of the models were estimated using classification of explanatory variables for each trip purpose
regression techniques that could produce somewhat surprising and thereby select dimensions across which the trip produc-
results. For example, regression model calibration techniques tion rates were categorized.
can result in negative coefficients for some of the variables. A The ANOVA results indicate that all of the independent
home-based shop trip attraction model could have, say, a posi- variables have significant effects on home-based work trip
tive coefficient for retail employment and a negative coefficient production rates. However, among all interaction effects,
for basic employment. Such occurrences might be explained as the household vehicles versus household workers variable
“second-level” relationships—each retail employee attracts a appears to be the strongest predictor of the home-based
certain number of home-based shop trips during the day, but work trip production rate. For home-based nonwork and
as the amount of basic employment increases around the retail home-based other trips, household workers versus house-
location, the number of home-based shop trips decreases due hold persons appears to be the strongest predictor of the trip
to unattractiveness of, say, an industrial area. production rate. For the nonhome-based trip purpose, the
However, some illogical regression results were also observed ANOVA results suggest that household workers by house-
in the review. An example is a home-based work model using hold persons is again found to be the strongest predictor of
multiple employment categories as independent variables the trip production rate.
with some of the coefficients being positive and some nega- The MPO Documentation Database indicated that two
tive. Since each employee should attract a reasonable average other cross-classifications are commonly used: number of
number of home-based work trips each day, a negative model persons by income level and number of persons by number
coefficient for an employment category is not logical. of vehicles. Parameters for these cross-classifications, also
estimated from the NHTS data set, are presented for all trip
4.4.4  Model Parameters purposes.
For home-based school trips, trip rates were estimated for
Trip Productions the cross-classification of number of persons by number of
The household trip production rates classified by variables children. Since some modeling agencies do not forecast the
representing household characteristics were estimated from number of children, trip rates were also estimated for num-
the 2009 NHTS data. These rates represent the number of ber of persons by income level and number of persons by
person trips, including both motorized and nonmotorized number of vehicles.
trips, per household. To determine the best variables to use Tables C.5 through C.9 in Appendix C show the trip rates
for the rates provided here, trip rates were summarized for by purpose cross-classified by the preferred pairs of variables,
the following variables: based on 2009 NHTS data, for home-based work, home-
based nonwork, nonhome-based, home-based school, and
• Number of persons, home-based other trips, respectively. The NHTS data showed
• Number of workers, nearly the same trip rates for all population ranges for most
• Income level, and trip purposes, apparently due at least in part to the relatively
• Number of vehicles. low sample sizes and resulting large errors associated with
some of the cells. For home-based nonwork and home-
The number of persons categories ranged from 1 to 5+. The based other trips, the NHTS data indicated lower trip rates
number of workers categories ranged from 0 to 3+. The num- for urban areas under 500,000 in population and nonurban
ber of vehicles categories ranged from 0 to 3+. The household areas, and so separate rates are presented for such areas for
income levels (in 2008 dollars) were defined as: these trip purposes.
Use of a cross-classification trip production model requires
• $0 to $9,999; that the households in each zone are classified along the same
• $10,000 to $24,999; dimensions as the model. For example, if the first model in

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

41  

Table 4.2.  Example number of households by numbers of persons and autos.

Persons
Autos 1 2 3 4 5+ Total
0 10 10 10 0 0 30
1 50 100 70 20 10 250
2 0 150 200 100 50 500
3+ 0 0 40 80 100 220
Total 60 260 320 200 160 1,000

Table C.5 is used, the households in each zone must be cross- by the NHTS, as represented by Tables C.5 through C.7, indi-
classified by number of workers (0, 1, 2, and 3+) and number cate that the average household in urban areas of greater than
of autos (0, 1, 2, and 3+). If the demographic estimates avail- 500,000 in population makes 10.0 person trips: 1.4 home-
able to the modeler are not already classified in the required based work trips, 5.6 home-based nonwork trips, and 3.0
manner, there are procedures that may be used to estimate nonhome-based trips. The average household in urban areas
the percentages in each cell and to apply them to the total of less than 500,000 in population makes 9.5 person trips:
households. Common sources for these percentages include 1.4 home-based work trips, 5.1 home-based nonwork trips,
the CTPP, NHTS, and local survey data. Depending on sam- and 3.0 nonhome-based trips. The range of person trips per
ple sizes, however, these sources may not provide statistically household in the MPO Documentation Database is about 1.3
significant percentages at the zone level, and it may be nec- to 2.0 home-based work trips, 2.6 to 5.9 home-based non-
essary to estimate percentages for groups of zones based on work trips, and 1.6 to 4.5 nonhome-based trips. Total person
area type and location within the region. trips per household range from 7.0 to 11.5.

Example Calculations Trip Attractions

Consider a zone with 1,000 households located in an urban Table 4.4 summarizes average daily trip attraction rates
area of under 500,000 in population where a trip production for the classic three trip purposes from the analyses of the
model with the classic three trip purposes is being developed. models in the MPO Documentation Database. These rates
The MPO has estimated the number of households in the were all estimated from local or statewide household travel
zone cross-classified by number of persons and number of surveys. While all of these models used person trips as the
vehicles, as depicted in Table 4.2. unit of travel, some used person trips in motorized modes
For home-based work trips, the number of households in while others used total person trips, including those by walk-
each cell is multiplied by the trip rate from the second section ing and bicycling.
of Table C.5, yielding the number of home-based work trips While Table 4.4 shows average rates for commonly defined
in each cell of the cross-classification in Table 4.3. models, achieving commonality required substantial process-
So this zone produces 1,839 home-based work trips. Simi- ing. Although trip attraction models are defined for the clas-
larly, home-based nonwork and nonhome-based trip produc- sic three trip purposes, development of rates for home-based
tions can be computed using the fourth section of Table C.6 nonwork and nonhome-based trips often required aggrega-
and the second section of Table C.7, performing the same type tion of more purpose-specific submodels. For example, if a
of calculations. region used both home-based shop and home-based other
Reasonableness checks of the trips per household by pur- (representing nonwork and nonshopping travel) trip attrac-
pose estimated from trip production model results can be tion models, the trip rates per retail employee were added in
performed. Information on the national sample represented the composite home-based other trip attraction model. If

Table 4.3.  Example number of home-based work trips.

Persons
Autos 1 2 3 4 5+ Total
0 2 7 11 0 0 20
1 30 80 84 34 15 243
2 0 195 400 200 115 910
3+ 0 0 104 232 330 666
Total 32 282 599 466 460 1,839

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

42

Table 4.4.  Trip attraction rates from selected MPOs (person trips per unit).

Number of Employment
MPO Models School
Summarized Householdsa Enrollmentb Basicc Retaild Servicee Total
All Person Trips
Home-Based Work
Model 1 16 1.2
Home-Based Nonwork
Model 1 2 1.2 1.4 0.2 8.1 1.5
Model 2 8 2.4 1.1 7.7 0.7
Model 3 2 0.7 0.7 8.4 3.5
Nonhome Based
Model 1 5 0.6 0.5 4.7 1.4
Model 2 8 1.4 6.9 0.9

Motorized Person Trips

Home-Based Work
Model 1 8 1.2
Home-Based Nonwork
Model 1 1 0.4 1.1 0.6 4.4 2.5
Model 3 4 1.0 0.3 5.9 2.3
Nonhome Based
Model 1 6 0.6 0.7 2.6 1.0
a
The number of households in a zone.
b
The number of elementary, high school, or college/university students in a zone.
c
Employment primarily in two-digit North American Industry Classification System (NAICS) codes 1–42 and 48–51
[Standard Industrial Classification (SIC) codes 1–51].
d
Employment primarily in two-digit NAICS codes 44–45 (SIC codes 52–59).
e
Employment primarily in two-digit NAICS codes 52–92 (SIC codes 60–97).
Source: MPO Documentation Database.

a region stratified trip attraction rates by area type, aver- 20 households, no school enrollment, 200 basic employees,
ages of the trip rates were estimated. If data were available 10 retail employees, and 100 service employees, the home-
for the various strata that had to be combined, weighted based nonwork trip attractions computed from Model 3 are:
averages were estimated; where data were not available 0.7  20 + 0.7  200 + 8.4  10 + 3.5  100 = 588.
for weighted averages, simple averages were used. Finally, Table 4.4 shows substantial variation in the trip attrac-
composite trip rates were estimated for three main employ- tion rates for the various model forms. The variation may
ment groups: basic employment, retail employment, and reflect the different sizes of urban areas, different travel
service employment. characteristics, and different development densities or
Since the presence or absence of other variables in a model area types, as well as the impact of variables included or
can affect the coefficient for a specific model variable, Table 4.4 excluded from the different model forms. It should be
shows sets of trip rates for trip attraction models with com- noted that no trends in trip attraction models by urban
mon independent variables. Rates are provided for all per- area population were evident; although the number of
son trips and motorized person trips only. Note that there models examined is small, this is consistent with previous
are some combinations of variables that none of the models documentation efforts such as NCHRP Report 365 (Martin
in the database used for motorized person trip attraction and McGuckin, 1998).
models. The trip attraction rates shown in Table 4.4 may provide
To use the information in Table 4.4 to obtain param- reasonable starting points for models for areas lacking the
eters for trip attraction models, the analyst should choose a locally collected data necessary to develop trip attraction
model that is consistent with the unit of travel (motorized or models. The selection of the specific model forms to be used
nonmotorized trips) and variables that are available for use could be made based on the types of independent data avail-
in model application. The number of attractions can then able for model application. The results of such initial model
be computed for each zone. For example, for a zone with specifications should be reviewed to ensure that they reflect

Copyright National Academy of Sciences. All rights reserved.

Travel Demand Forecasting: Parameters and Techniques

43  

known travel conditions and behave reasonably for a region. models are multivariate. The coefficients of a model that has
Three examples are provided in the following paragraphs. the same variables could be compared to those in one of the
These examples all use the models for “all person trips” in the models in Table 4.4, but having the same or different coef-
upper portion of Table 4.4. ficients as one other model would not provide confirmation
of the reasonableness or unreasonableness of the model.
Example 1.   Suppose the trip attraction rates from home- For home-based work trips, the vast majority of attraction
based work model 1, home-based nonwork model 3, and models in the MPO Documentation Database have coef-
nonhome-based model 1 are applied for a region. In a review ficients for total employment in the range of 1.0 to 1.5, and
of traffic assignment results, it is discovered that too many so coefficients in this range may be considered reasonable.
trips are crossing the cordon boundary around the CBD. In
such a case, it might be reasonable to reduce the home-based
4.5  Trip Distribution
nonwork and nonhome-based trip attraction rates for retail
and service employment in the CBD and to balance those Trip distribution is the second step in the four-step mod-
reductions in the CBD trip rates with increases of the values eling process. It is intended to address the question of how
for the rates for non-CBD zones. However, before making many of the trips generated in the trip generation step travel
such adjustments, other checks should be performed, includ- between units of geography, e.g., traffic analysis zones. These
ing the accuracy of CBD socioeconomic data, mode shares to trips are in the same units used by the trip generation step
the CBD, and comparison of CBD through traffic to observed (e.g., vehicle trips, person trips in motorized modes, or per-
origin-destination data. son trips by all modes including both motorized and non-
motorized modes). Trip distribution requires explanatory
Example 2.   Suppose a region has forecasts for only variables that are related to the impedance6 (generally a func-
households, retail employment, and nonretail employ- tion of travel time and/or cost) of travel between zones, as
ment available. None of the three home-based nonwork well as the amount of trip-making activity in both the origin
model forms match the independent variables available for zone and the destination zone.
the region. In this case, it might be reasonable to test both The inputs to trip distribution models include the trip
home-based nonwork models 2 and 3, ignoring the coeffi- generation outputs—the productions and attractions by
cients for the missing variables. Careful attention should be trip purpose for each zone—and measures of travel imped-
paid to traffic assignment results around industrial areas and ance between each pair of zones, obtained from the trans-
educational facilities. The “best performing” model in terms portation networks. Socioeconomic and area characteristics
of reproducing traffic volumes would be selected. If neither are sometimes also used as inputs. The outputs are trip
model performed well, it might be appropriate to mix the tables, production zone to attraction zone, for each trip
rates to address the issues. purpose. Because trips of different purposes have different
levels of sensitivity to travel time and cost, trip distribution
Example 3.   Again, suppose a region has employment is applied separately for each trip purpose, with different
stratified only by retail and nonretail at the zone level. If model parameters.
regional totals for basic and service employment can be
determined, nonretail attraction rates for the home-based
4.5.1  Model Function
nonwork and nonhome-based trip purposes can be esti­
mated by applying home-based nonwork model 1 (or model 3) The gravity model is the most common type of trip distri-
at the regional level and estimating a weighted average trip bution model used in four-step models. In Equation 4-9, the
rate for nonretail employment. The same procedure could denominator is a summation that is needed to normalize the
be applied using rates from nonhome-based model 1 to gravity distribution to one destination relative to all possible
develop a weighted average nonretail employment trip rate. destinations. This is called a “doubly constrained” model
If the regional totals for basic and service employment are because it requires that the output trip table be balanced to
not available, the straight averages of the rates for basic and attractions, while the numerator already ensures that it is bal-
service employment could be used. For example, if using anced to productions.
model 3 for home-based nonwork attractions for motor-
ized trips, one could use the average of the basic and service
employment coefficients (1.3) as the coefficient for nonretail 6
The term “impedance” is used in this report to represent the general-
employment. ized cost of travel between two zones. In most cases, the primary com-
It is difficult to perform reasonableness checks of trip ponent of generalized cost is travel time, and so impedance is often
attraction model results for most trip purposes because the expressed in time units such as minutes.

Copyright National Academy of Sciences. All rights reserved.