Use of Accessibility Measure in Trip Generation Models

• As no transportation system variable is used in trip generation analysis, changes in the

transportation system (such as introduction of new high speed road link, metro rail link, etc.) have
no effect on productions and attractions.
• To overcome this disadvantage, modellers have used accessibility measure as one of the
explanatory variables.
• Accessibility measure is directly related to the opportunities available and inversely related to the
deterrence in reaching these opportunities.

𝐴𝑀𝑖 = ෍ 𝑓(𝐸𝑗 , 𝐶𝑖𝑗 )

𝑗
Where, 𝐴𝑀𝑖 = Accessibility Measure
𝐸𝑗 = Attraction (opportunities, such as employment) of zone j
𝐶𝑖𝑗 = Generalised cost of travel between zone i and zone j = 𝛼 𝑇𝑖𝑚𝑒 + 𝛽 𝐶𝑜𝑠𝑡
A typical accessibility measure that is generally used is:
𝐴𝑀𝑖 = ෍ 𝐸𝑗 𝑒 −𝛽𝐶𝑖𝑗
𝑗
Where, 𝛽 = calibration parameter
 Generalised Cost of Travel
𝑐𝑖𝑗 = 𝑎1 𝐼𝑉𝑇𝑇𝑖𝑗+𝑎2 𝑊𝐾𝑇𝑖𝑗 + 𝑎3 𝑊𝑇𝑖𝑗 + 𝑎4 𝑇𝑅𝑇𝑖𝑗 + 𝑎5 𝐹𝐴𝑅𝐸𝑖𝑗 + 𝑎6 𝑇𝐶𝑗 + α
Where, 𝐼𝑉𝑇𝑇𝑖𝑗 = in vehicle travel time between i and j
𝑊𝐾𝑇𝑖𝑗 = walking time of the trip between i and j
𝑊𝑇𝑖𝑗 = waiting time for the trip between i and j
𝑇𝑅𝑇𝑖𝑗 = transfer/interchange time for the trip between i and j
𝐹𝐴𝑅𝐸𝑖𝑗 = perceived cost of travel for the trip between i and j.
This is taken as fare paid if mode is transit or taxi or cost of fuel if mode is private vehicle.
𝑇𝐶 j = cost incurred at the transport terminal (such as parking fee)
α = Unaccounted costs such as comfort, convenience, etc.
𝑎1 , 𝑎2 , 𝑎3 , 𝑎4 , 𝑎5 and 𝑎6 are the weights for converting each cost component
into common unit i.e., either money (fare) or time (in vehicle travel time) units.
If 𝑎5 = 1, then 𝑎1 , 𝑎2 , 𝑎3 , 𝑎4 are interpreted as values of in vehicle travel time,
walking time, waiting time and transfer time respectively. If people perceive
parking charges as more taxing than fare, 𝑎6 will be more than 1.
 Trip Distribution
• The purpose of trip distribution phase is to synthesise trip linkages
between zones.
• i.e., to determine how the trips produced in one zone are distributed
to all other zones
• Transport planners are basically interested in determining the
destination choice of the travellers.
• Generally the traffic distribution is influenced by
• Nature, type and characteristic of transportation system
• Magnitude and spatial distribution of land use
• Socioeconomic characteristics of the population
 Origin – Destination Matrix: Notation
Destinations
Origins 1 2 3 ….j …n
෍ 𝑇𝑖𝑗
𝑗

1 T11 T12 T13 …T1j …T1n O1

2 T21 T22 T23 …T2j …T2n O2
3 T31 T32 T33 …T3j …T3n O3
.
.
.
i Ti1 Ti2 Ti3 …Tij …Tin Oi
.
.
.

n Tn1 Tn2 Tn3 …Tnj …Tnn On

෍ 𝑇𝑖𝑗 D1 D2 D3 …Dj …Dn ෍ 𝑇𝑖𝑗 = 𝑇

𝑖 𝑖𝑗
 Notation Contd.
Tij = trips made between zone i and zone j
Notations in lower case letters will be
Oi = trips originating in zone i = ෍ 𝑇𝑖𝑗 used to denote base year or prior trip
𝑗 matrices
Dj = trips having destination at zone j = ෍ 𝑇𝑖𝑗
𝑖

T = total number of trips = ෍ 𝑇𝑖𝑗

𝑖𝑗
𝑝𝑚
𝑇𝑖𝑗 = trips made between zone i and zone j for purpose p by mode m

Pi = Productions of zone i

Qj = Attractions of zone j
𝑚
𝑐𝑖𝑗 = Generalised cost of travel between zone i and zone j for mode m
 Growth Factor Methods
• These methods simply use the growth rate in trips either for the
whole of the study area or separately for each zone in terms of
productions and attractions, for getting the trip distribution for future
years
• Following are the different growth factor methods used depending on
the extent of information available on growth factors:
• Uniform growth factor methods
• Singly constrained growth factor methods
• Doubly constrained growth factor methods
 Uniform Growth Factor Method
In this method, as the only information available is the overall growth
rate of trips for the whole of the study area, each cell of the trip matrix is
multiplied with this single growth factor.
𝑇𝑖𝑗 = 𝜏. 𝑡𝑖𝑗 , for each pair of i and j
𝑇𝑖𝑗 = Future trips from zone i to zone j
𝑡𝑖𝑗 = Present or base year trips from zone i to zone j
𝜏 = general growth rate for all zones
𝜏 =T/ t (Future year total trips/base year total trips)
 Example on Uniform Growth Factor Method
Base year trip matrix Future year trip matrix with  = 1.2

1 2 3 4 oi 1 2 3 4 Oi

1 5 50 100 200 355 1 6 60 120 240 426

2 50 5 100 300 455 2 60 6 120 360 546

3 50 100 5 100 255 3 60 120 6 120 306

4 100 200 250 20 570 4 120 240 300 24 684

dj 205 355 455 620 1635 Dj 246 426 546 744 1962

As the growth of traffic in the study area is expected to be 20% over the
forecasting period, each cell of the O-D matrix is multiplied by 1.20
 Singly Constrained Growth Factor Methods
• When information is available on the expected growth in either trips originating in
each zone or trips attracted to each zone, this method comes into picture
𝑇𝑖𝑗 = 𝜏𝑖 . 𝑡𝑖𝑗 , for origin/production specific factors
𝑇𝑖𝑗 = 𝜏𝑗 . 𝑡𝑖𝑗 , for destination/attraction specific factors
Where, 𝜏𝑖 = growth factor for origin i = Oi/oi
𝜏𝑗 = growth factor for destination j = Dj/dj

The above method is useful when either of the following situations happen:
• Only productions can be computed accurately for the zones as the corresponding
zonal explanatory variables (e.g., population) can be estimated accurately
• Only attractions can be computed accurately for the zones as the corresponding
employment and shopping variables can be estimated accurately
 Example on Singly Constrained Growth Factor Method
Base year origin constrained trip matrix Future year origin-constrained trip matrix
Target
1 2 3 4 oi i 1 2 3 4 Oi
Oi

1 5 50 100 200 355 400 1.13 1 5.6 56.3 112.7 225.4 400

2 50 5 100 300 455 460 1.01 2 50.5 5.1 101.1 303.3 460

3 50 100 5 100 255 400 1.57 3 78.4 156.9 7.8 156.9 400

4 100 200 250 20 570 702 1.23 4 123.2 246.3 307.9 24.6 702

dj 205 355 455 620 1635 1962 Dj 257.8 464.6 529.5 710.1 1962

I = Oi /oi
 Doubly Constrained Growth Factor Methods
• These methods apply when information on both productions and attractions (or
origins and destinations) is available, i.e.,
i = Origin/production specific growth factor and
j = Destination/attraction specific growth factor are available.
• These growth factors are applied alternatively in an iterative fashion, such that
both the trip end constraints are satisfied, i.e.,
• Oi = ෍ 𝑇𝑖𝑗
𝑗

• Dj = ෍ 𝑇𝑖𝑗
𝑖
 Doubly Constrained Growth Factor Methods

• Several such iterative methods were proposed in the literature. In all these
methods, intermediate correction factors are worked out after every iteration by
comparing the row/column totals with the respective target trip ends. These are
listed below:
• Average factor method
• Detroit method
• Fratar method
• Furness method
 Furness (1965) Method
Furness introduced balancing factors for matching the trip end totals:
Tij = tij . i . j. Ai . Bj
Where, Ai . Bj are balancing factors
Incorporating these growth factors and balancing factors into new variables,
Tij = tij . ai . bj
Where,
ai = i . Ai
bj = j . Bj
The factors ai and bj are computed satisfying the trip end constraints as per the
following iterative process
Furness, K. P., (1965). “Time function iteration”, Traffic Engineering & Control, 7, 458-460.
 Iterative Procedure for Furness Method
Step 1: Compute ai for the first iteration as
ai=Oi/oi
Set all bj=1 and apply the factors ai to the trip matrix, satisfying the
origin/production constraint:
i.e., Tij= tij . ai
Step 2: Compute new bj as , bj = Dj/dj and apply the facors bj to the trip
matrix satisfying the destinations/attractions constraint:
Tij= tij . bj
Step 3: Compute new ai as ai = Oi/oi and apply the new ai factors to the trip
matrix satisfying the origin/production constraint.
Repeat steps 2 and 3 until the difference between the target and the computed
trip ends is sufficiently small.
 Furness Method – Iteration 1 with ai
Base year doubly constrained trip matrix Intermediate trip matrix with balanced origins
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi ai

1 5 50 100 200 355 400 1.13 1 5.6 56.3 112.7 225.4 400 400 1

2 50 5 100 300 455 460 1.01 2 50.5 5.1 101.1 303.3 460 460 1

3 50 100 5 100 255 400 1.57 3 78.4 156.9 7.8 156.9 400 400 1

4 100 200 250 20 570 702 1.23 4 123.2 246.3 307.9 24.6 702 702 1
dj 205 355 455 620 1635 dj
257.8 464.6 529.5 710.1 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 1.01 0.86 0.94 1.13

 Furness Method – Iteration 1 with bi
Base year doubly constrained trip matrix Intermediate trip matrix with balanced destinations
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi ai

1 5 50 100 200 355 400 1.13 1 5.7 48.5 106.4 254.5 415.1 400 0.96

2 50 5 100 300 455 460 1.01 2 51.0 4.4 95.5 342.5 493.3 460 0.93

3 50 100 5 100 255 400 1.57 3 79.1 135.1 7.4 177.2 398.7 400 1.00

4 100 200 250 20 570 702 1.23 4 124.2 212.1 290.7 27.8 654.9 702 1.07
dj 205 355 455 620 1635 dj 260 400 500 802 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 1 1 1 1

 Furness Method – Iteration 2 with ai
Base year doubly constrained trip matrix Intermediate trip matrix with balanced origins
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi ai

1 5 50 100 200 355 400 1.13 1 5.5 46.7 102.5 245.3 400 400 1

2 50 5 100 300 455 460 1.01 2 47.5 4.1 89.0 319.4 460 460 1

3 50 100 5 100 255 400 1.57 3 79.4 135.5 7.4 177.7 400 400 1

4 100 200 250 20 570 702 1.23 4 133.2 227.3 311.7 29.8 702 702 1
dj 205 355 455 620 1635 dj 265.5 413.6 510.6 772.2 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 0.98 0.97 0.98 1.04

 Furness Method – Iteration 2 with bi
Base year doubly constrained trip matrix Intermediate trip matrix with balanced destinations
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi ai

1 5 50 100 200 355 400 1.13 1 5.4 45.2 100.4 254.7 405.7 400 0.99

2 50 5 100 300 455 460 1.01 2 46.5 3.9 87.2 331.7 469.4 460 0.98

3 50 100 5 100 255 400 1.57 3 77.7 131.0 7.3 184.6 400.6 400 1.00

4 100 200 250 20 570 702 1.23 4 130.4 219.9 305.2 31.0 686.4 702 1.02
dj 205 355 455 620 1635 dj 260 400 500 802 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 1 1 1 1

 Furness Method – Iteration 3 with ai
Base year doubly constrained trip matrix Intermediate trip matrix with balanced origins
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi Ai

1 5 50 100 200 355 400 1.13 1 5.3 44.6 99.0 251.2 400 400 1

2 50 5 100 300 455 460 1.01 2 45.6 3.8 85.4 325.1 460 460 1

3 50 100 5 100 255 400 1.57 3 77.6 130.8 7.3 184.3 400 400 1

4 100 200 250 20 570 702 1.23 4 133.4 224.9 312.1 31.7 702 702 1
dj 205 355 455 620 1635 dj 261.9 404.1 503.8 792.3 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 0.99 0.99 0.99 1.01

 Furness Method – Iteration 3 with bi
Base year doubly constrained trip matrix Future Year fully balanced trip matrix
Target Target
1 2 3 4 oi Oi ai 1 2 3 4 oi Oi Ai

1 5 50 100 200 355 400 1.13 1 5.2 44.1 98.2 254.2 401.9 400 1.00

2 50 5 100 300 455 460 1.01 2 45.3 3.8 84.8 329.1 463.0 460 0.99

3 50 100 5 100 255 400 1.57 3 77.0 129.5 7.2 186.6 400.3 400 1.00

4 100 200 250 20 570 702 1.23 4 132.4 222.6 309.8 32.1 696.8 702 1.01
dj 205 355 455 620 1635 dj 260 400 500 802 1962
Target Target
260 400 500 802 1962 260 400 500 802 1962
Dj Dj

bi 1.27 1.13 1.10 1.29 bi 1 1 1 1

 Limitations of Growth Factor Methods
• Any error in the base-year data gets amplified by the iterative
application of the correction factors
• Can not be used to forecast unobserved cells of base-year trip
matrices
• Does not take into account changes in transport costs due to
improvements in the network
• Can not be used in analysis of policy options involving new modes,
new links, pricing policies and new zones