CE 457

Urban Transportation Planning

and Management
2.00 credit, 2 hrs/week

Introduction to Demand Modelling-

Transportation System Analysis

By Dr. Md. Hadiuzzaman, Professor, Dept. of CE, BUET

System - A Definition
A system is a group of interdependent and interrelated
components that form a complex and unified whole
intended to serve some purpose through the performance
of its interacting parts

complex

unified

purposeful

whole more than the sum of the parts

3

Traffic System

road network vehicles signalling

A Systems Perspective of
Transportation

System Concepts
• Hierarchy
• Purpose
• Boundary
• Components
• Performance
• Capacity
• Control
• Feedback
Transportation in a Systems
Hierarchy
Transportation System Purpose
•Provide mobility for the movement of goods & persons

•Provide accessibility to economic & social activities

•Promote sustainable urban form

•Enable/facilitate a broad range of economic & social

objectives

•Major economic sector in its own right

•….

All in an equitable, efficient, effective, safe manner

System Boundary

Where to draw the

boundary?

How to connect with

the rest of the world?

Boundary effects in
the analysis?
 The systems boundary can be small or large, depending on the purpose of
analysis, and the expected scale of the impacts
 E.g. proposed bike-lane on Burrard bridge may cause traffic to spill over
onto parallel roads, but it is not necessary to model the entire city in detail
GVRD Model External
Zone System

 E.g. Many large scale infrastructure improvements require that we model a

much larger area in detail. E.g. implementing HOV lanes on Hwy 1 would
have broad impacts on trips across the city
Performance & Capacity
Not the primary focus of this course.

Other courses deal in detail with this issues

CE 351: Transportation Engineering I; …)

We will take a “planning perspective”concerning

performance & capacity.
Performance, cont’d
Primary interests in this course:
• Extent to which system is achieving its
goals/objectives/standards

• Assessment of “non-transportation” system impacts (i.e.,

the transportation system generates many more “outputs”
than just the movement of people & goods)
System Control
• Operational control (not the focus of this course)

• “Institutional” control
• policies
• regulations
• provision of services
• provision of infrastructure
• ….

Focus of the planning process

System Control through Policy
Policy Tools:
• fiscal
• regulatory
• operational
Community • infrastructure
Goals • marketing
• education
• information
Demand Management dissemination

The purpose of systems analysis is to assess the impacts of

proposed policies on transportation system performance,
and on society as a whole.
Representing the Urban System -
Characteristics of Urban Travel
Travel occurs in space and time.

Travel occurs so that people can participate in activities -

-> generated trips intended to accomplish different
purposes.

Travel is executed via particular modes of travel, which

use specific routes through the (multimodal)
transportation network.
Measures of Travel Demand
• Vehicle (or passenger) kilometres travelled
• Trip chains (# of times a person leaves the house)
• Trips
• ….

The trip is the basic unit of most travel demand

analyses.

A trip is the movement from

a single origin to a single
H W
destination for a single
purpose
What is trip chaining?
Dimensions of Travel Demand
To fully characterize travel demand (trips), we need to
determine:

• purpose
• time of day
• origin
• destination
• mode(s)
• route
• frequency
Representing Space & Activities -
Zones
The urban area is divided up into a set of mutually exclusive,
collectively exhaustive zones.

Zones contain people, jobs, buildings, activities, etc.

Zone boundaries usually are defined by major streets,

geographic features (rivers, etc.), political boundaries, etc.

It is useful to make zones as homogeneous as possible, but this is

difficult to do in practice
Example of a Zone System
Transportation Network
Representation: Links & Nodes
The transportation network is represented by a connected
set of links and nodes

Links generally correspond to major streets, transit lines,

etc. Nodes generally correspond to intersections, transit
stops or stations, etc

Links and Nodes both have attributes (link speeds,

capacities, etc.; node transit transfer times, node roadway
turning restrictions, etc.).

The computer representation of the network is always a

simplification and abstraction of the real network
Model
Example Network
LEGEND
Provincial Highways
Major Road Network
Airport Authority
Linking Space & Networks -
Centroids, Centroid Connectors
Centroids are special, artificial nodes which represent
zones (i.e., one centroid exists for each zone).

Centroid connectors are special, artificial links which

connect centroids to “real” transportation network nodes.

Use of centroids effectively assumes that

all activity (people, jobs, etc.) is located
at the centroid.
i.e., all trips to/from the zone are treated
as if they are to/from the zone centroid.
Example of Centroid & Centroid
Connectors
Representing Time
 Travel demand changes depending on the time
of day

 Many planning applications require that we

break down the day into
 AM Peak period (6:00am – 8:59am)
 Midday period (9:00am – 2:59pm)
 PM Peak period (3:00pm - 5:59pm)
 Off peak (6:00pm – 5:59am)
Process of Travel Demand Modelling

the demand Data Collection Choose Modelling

Method
Base Demand (D)
analysis Model Estimation

process
& Calibration
Base Attributes

Other Historical Model

Data Validation

Forecasts of
ex Variables
Forecast Demand/
Analyze Policies
Specify Policy
Options
Model Estimation
All models involve trying to relate one or more
dependent or endogenous variables to one or more
independent or exogenous or explanatory variables:

D = f(X,q) [1]

D = demand (dependent variables)

X = explanatory variable(s)
q = parameter(s)
Model Estimation
All models involve one or more parameters, which reflect
tastes, preferences, weights, trade-offs between variables
in determining the behaviour (dependent variable(s))
being modelled

In hypothesizing the model, we have some idea about

what factors influence behaviour, but we need to quantify
that influence

This is the process of model estimation. We use observed

data to estimate model parameters.
Model Estimation
Trips (T)

Single data points, (Ti,Xi)

Income (X)

Hypothesized model: T = a + bX q = {a,b}

Model Estimation Problem

Find the values of {a,b} which define the straight line which “best fits”
observed data for the system being modelled
Model Calibration
After a model’s parameters are statistically estimated, it
sometimes is necessary to “calibrate” the model by
adjusting these parameters (or adding additional
parameters to the model) in order to force the model to fit
base case data to a pre-specified level of accuracy

E.g. adjust the y-intercept so that we correctly replicate

base year road counts
Model Validation
Once a model has been estimated/calibrated to base year
conditions, its predictive performance should be tested
or “validated” by running the model for another
situation (e.g., another year) which was not used in the
model development, to see how well the model performs
in predicting this other, known system outcome.

Usually, the same goodness-of-fit criteria used to

estimate the model can be used to assess model
performance in forecasting.
Trend Analysis
The simplest form of demand model involves
extrapolating historical trends into the future

This can be useful as a diagnosis of system performance,

but it does not provide a mechanism for forecasting how
travel demand might change in response to policy
changes

It is also difficult to choose the correct functional form for

extrapolation
The Urban Transportation
Modelling System (UTMS)
To forecast urban travel demand, a standard modelling
framework has been developed over the past 50 years (!).
This is the four-step/FSM or UTMS modelling framework.

Many variations of this framework exist, but the general

paradigm is used world-wide for operational travel demand
forecasting
UTMS - Key Characteristics
Key, distinguishing characteristics of the UTMS approach
to travel demand modelling include:
• Zone-based
• Trip-based
• Aggregate time-periods
• Limited socio-economic attributes of trip-makers
• Multiple trip purposes (separate models for each
purpose)
Definitions
N = Number of zones in the urban area
i = Subscript, used to denote origin zones
j = Subscript, used to denote destination zones
Oi = Number of trips originating in zone i
Dj = Number of trips destined for zone j
Tij = Number of trips (“flow”) from origin zone i
to destination zone j
Tijm = Number of trips from i to j using mode m

(for a given trip purpose, p)

4 Stages of Travel Demand Modelling

I Trip Generation J

Oi Dj
Trip Distribution
I J
Tij

Tij,auto

I Mode Split J

Tij,transit
Four-Stage System, cont’d

Trip Assignment
-- path of flow Tij,auto
through the auto
network
The Urban Transportation Modelling
System (UTMS)
Population & Employment
Forecasts

Trip Generation

Trip Distribution

Transportation
Mode Split Network & Service
Attributes

Trip Assignment

Link & O-D Flows,

Times, Costs, Etc.
Multiple Trip Purposes/Time Periods
Previous flow chart applies for a single trip purpose, for a
single time period

To model multiple purposes, must do trip generation,

distribution & mode split for each purpose

To model multiple time periods, must do all four stages for

each period
Example -
3 Trip Purposes, One Time Period
Population Employment

HW HS NWS

Generation Generation Generation

Trip Rates,
etc.
Distribution Distribution Distribution

Mode Split Mode Split Mode Split

Transport
Network

Road Transit
Assignment Assignment

Link & O-D volumes, times, costs, v/c ratios, etc.