Addis Ababa Institute of Technology (AAiT), 2015

1 Transport Planning – CENG 6303

Introduction to
Transport
Modeling

Instructor: Dr. Bikila Teklu

bikilatek@yahoo.com
 Introduction
2

A model …
 isa simplified representation of the world or
the representation of an object or system in
another form.
 ranging from a few simple equations to
suites of computer programs.
 Introduction
3

Models can be..

 quantitative (with numbers) or qualitative (without)

 static (relating to states or conditions) or dynamic (relating to

processes)
 causal (why something happens) or simply correlative

 aggregate or disaggregate

 transparent (to aid understanding or explanation) or a black box

 predictive (what will happen) or prescriptive (what should happen)

 empirical (based directly on data) or synthetic (based on a

relationship derived from the data)
 Introduction
4

Advantages..
 help visualize and understand problems.
 important part of decision-making processes, allowing users to
explore and estimate the consequences of particular policies,
strategies or schemes on a desktop rather than in a real network.
 facilitate the discussion of assumptions and are (can be),
therefore, more transparent than mental models.
 assist in the analysis of risk and uncertainty.
 models can provide important inputs to the appraisal process
which help foster the efficient and transparent allocation of
scarce resources.
 Introduction
5

purpose of a model..
 to help understand how the system works
 to help explain or communicate how the system works
 to help predict usage and performance of the system
in various possible future circumstances
 to help design or manage facilities and services
 to help evaluate options for possible investment
 Introduction
6

level of detail and accuracy required..

 Simulation at different levels, e.g. micro, macro…
 Detail and precision may depend on these levels.
 Unnecessary detail may:
 hinder understanding.
 usually represents a waste of resources.
 A detailed model is more expensive but not necessarily a
better model, not even more accurate.
 However, disaggregation of data may reveal trends and
relationships not apparent at the aggregate level.
 Introduction
7

Simple model example

 Oh = a + bEh + cCh
 Where:
 Oh = number of trips made by household h
 E h = number of workers in household h
 Ch = number of cars owned by household h
 Oh is the dependent variable

 Eh and Ch are independent or explanatory variables

 a, b and c are parameters

 Introduction
8

Model specification, calibration and validation

 Model specification:
 Independent or explanatory variables
 Dependent or response variable
 Form of relationship (e.g. linear, exponent etc)
 Model calibration:
 Quantify parameters using observed data
 Model validation:
 Consider other situations by using different data from
different locations or time periods
 Introduction
9

Modeling errors
1. Specification
2. Data
3. Calibration
4. Aggregation
5. Transfer
Important to note:

“All models are wrong, but some are more
wrong than others”
 Introduction
10

Uncertainty in Models
Uncertainty / error

Model-complexity
 Transport Models
11

What is a transport model ?

 A simplified mathematical representation of a small part
of the real world, aiming at describing and explaining
travel behavior and visualizing the amount and patterns of
transport.
 Mathematical models which on basis of data for land use
and the transport system, together with an understanding
of human behavior, calculate the resulting traffic on the
transport network.
 For the present situation, these calculations can be compared
with the traffic we can observe on the roads, in the streets and
on the public transport system.
 The Development and Short History of
Transport Models
12

1960 – 1975: The first complete models in use; USA and Europe
Driving forces:
The increasing traffic
Theoretical development
The computers (important prerequisite)
Focus:
Traffic growth / forecasts; How large roads do we need?

1975 – 1985: Models got a bad reputation in many countries, focus

was now on traffic safety and environmental condition – but still research
in the universities – theoretical development

1985 – 1995: Revival of the models: but focus on getting traffic data
to calculate safety and environmental effects of different projects / traffic
plans
 The Use of Transport Models Today ?
13

Providing traffic data for the future; thus understanding the

challenges to come; - - - congestion, environmental
conditions, future transport costs for people and for
business/industry, accessibility and mobility, etc

Analyzing the effect of alternative traffic plans/projects

Providing traffic data for “benefit-cost-analysis”

Providing traffic data for calculating the development of
traffic accidents and environmental conditions
 Definitions of networks & zones
14

Networks
 graphical representations of the transport system.
 usually represented as a set of zones and links.
 encompass large enough area to study all significant
impacts and influences.

Zones
 reduction of the trip matrix into manageable proportions.
 size of zones: compromise between accuracy and economy.
 boundaries chosen to reflect (1) zones for which data is
available, e.g. enumeration districts or political units
(2) zones used in previous studies.
 distinguish between “internal zones” in the study area
bounded by external cordon and “external zones” covering
the rest of the world.
 Network and matrix
15
 Ideal zones
16

Consistent with land use and population characteristics

Divided by distinct barriers to movement (e.g. rivers)

Centered around points of access to transport facilities
(e.g. stations)

Roughly equal size

“Hierarchical Zoning Systems”
 Typical zones
17

Geocoding zones – based on grid references (easy

to plot and aggregate but may have little regard to
the rules outlined above).

Postcode zones – valuable when collecting
questionnaire or interview data.

Centroids
 Gravitational centers of zones.
 Assuming that all trips from/to a given zone will start/end
at its center.
 Useful when calculating interzonal costs.
 Addis Ababa
18
 The basic principles of demand
19

Demand is the amount of a product or service desired

at a particular price in a given time period.

This is shown graphically in a demand curve or
numerically in a demand schedule.
 The basic principles of demand
20

Demand for transport (Qi) is typically taken to be a

function of:
 The price of the service (Pi)

 The price of other goods and services (Pn)

 The quality of the service (Si) (e.g. journey time, service

frequency, etc)
 The quality of other goods and services (Sn)

 Income (Y)

Qi = f(Pi,Pn,Si,Sn, Y)
 Equilibrium
21

Within a given time period, the transport system is

in a state of balance such that the demand has
adjusted to suit the supply and/or that the supply
has adjusted to suit the demand.

If demand exceeds capacity (e.g. too many cars

on a route causing delays), some drivers choose to
use other routes, and demand reduces.
 Transport Modeling
22

We organize the transport model in four basic stages:

1. Trip production and attraction: Models calculate the number
of trip produced and attracted to each zone in the study
area based on data for land use
2. Trip distribution: Models calculate the trip pattern
connecting trip productions and attractions
3. Mode choice: Models distribute trips from an origin (zone)
to a destination to the different modes of transport
4. Traffic assignment: Car trips are distributed to the road
network and public transport trips are distributed to the
public transport network
 Four stage modeling in a nutshell
23

1. Trip production 2. Zone-to-zone distribution

1
1

2 3

2 3

4 5 6
5
Population 4 6

Jobs
Trip ends

3. Mode choice 4. Traffic on roads and 1trips on public transit

1

2 3
2 3

5
Car
4
Public transit 6
Walk / Bike
5
4 6
 Four stage modeling in a nutshell
24

A transport model normally consists of four basic stages

(the four-stage-modeling):
1. The first step converts the land use to trip production
2. The second step calculates the Origin – Destination (OD) –
matrixes by using a measure of the spatial separation
between origin and destination and knowledge of how people
react to this separation (distance, time, cost etc)
3. In the third step, the model calculate how people choose
mode for their travel by using the characteristics of each
mode and car ownership as the most common explanatory
variables
4. In the last step, the model assigns car traffic to the road
network and the transit trips to the transit lines (shortest
route, capacity restraint)
 Four stage modeling in a nutshell
25

Stage ”In-Data”
Trip Production Land use
- Zonal population (age and sex)
- Zonal jobs (categories)
- Car ownership (0, 1 or 2+)
Trip distribution Trip production
Spatial separation (Distance, time, cost, - - )
between zones
Transport-standard
- Travel time (waiting inclusive)
- Travel cost
- Parking
Modal Choice Socioeconomic data
- Car ownership
- sex
Traffic Assignment Roads: link-distances, speed, traffic volumes,
- Car delay-functions
- Public Transport Public transport; scheduled times, frequency
and attachment to land use
 Organizing the Transport Model
26

The structure of the transport model will be

according to ”the four-stage-modeling”, but there
are some preparations to do before we start the
modeling work:
 Zoning and land use data
 Transport network coding
 Data acquisition (Inventories, travel surveys)
 Zoning and Land Use Data
27

The most important condition is the availability of zonal land use data
 Population; when we start to look into trip production models,
population data may be needed by “households”, “age-group”,
”income-group”, “sex” and “car ownership”
 Jobs; number of jobs in each zone, split into different type of jobs;
the simplest stratification is two groups of jobs:
 Ordinary jobs which do not attract customers or visitors, the only
trips made are the journey to and from work (and may be some
freight transport)
 Jobs which attract customers (public attractive jobs)
 Other land use influencing trip production; shopping malls, cinemas,
sport arenas, schools, universities, etc
 Zoning - Basic Principles
28

Inside a zone we define a ”centroid”, the zone’s trip

production center

The centroid is connected to the road network and to the
public transport system by connector links

The zonal system should be designed in such a way that the
route choice for trips starting or ending in the zone will be as
realistic as possible, thus it is important to know how the
transport model works when you design the zonal system

Although zoning is a simplification, large cities are divided
into more than 1000 zones
 Network Coding
29

The objective is to describe a transport network in a digital form

The network is identified by characteristics such as
 Nodes (often located at intersections)
 Links

The links are described by:
 Type of link
 Link distance
 Link speed
 Link travel time
 Link Capacity
 Link delay function
 Location of nodes (geographical coordinates)
 Trip Definitions
30

When we work with transport models we need a precise

definition of a trip:

Age group (what about people less than 7, less than 13,
less than 18 ?)

What is the trip producing end and what is the attracting

end ?

What about walking and biking trips ?

What about distance, is there a minimum distance ?

What about bus drivers and taxi drivers ?

What about trips made while working ?

 Transport Modeling and the
Development of Computers
31

In a large city, say with 2000 zones, each of the

zone-to-zone matrixes will be 4,000,000
elements.

In the model we calculate several of those in
addition to many other sub-model calculations.

The development of transport models is closely
tied to the development of computing capacity
 A National System of Transport Models
32

National /International

Regional

Local / City

Management / simulation model

 Transport Models -
Classification of Model Types
33

Management models / simulation models

Tactical models

Strategic models

Time horizon
 Transport Models as Part of a Larger
Modeling Package
34

Integrated Transport and Land Use Models

Integrated with GIS systems

Integrated with models for

Economic Analysis (CBA-studies)

Environmental consequences
 Noise
 Pollution

Accidents by degree of seriousness
 Computer Packages
35

And many more…………..