Understanding the N-Logit (Multinomial Logit) Model in Transportation Studies
Introduction
Transportation systems play a vital role in shaping the daily lives of people. Every traveller
must decide how to move from one place to another—whether to walk, use a bus, metro, car,
or any other mode. Understanding how people make these travel choices is one of the key goals
of transportation planning. To study and predict such decisions, researchers use discrete choice
models. One of the most commonly used models is the Multinomial Logit (MNL) model,
sometimes called the N-Logit when extended into a hierarchical or nested form.
The MNL model helps planners and researchers understand how various factors like
travel time, travel cost, comfort, frequency, and safety influence a person’s decision to choose
one mode over others. This understanding is essential for designing efficient, affordable, and
sustainable transport systems.
What Is the N-Logit (Multinomial Logit) Model?
The Multinomial Logit (MNL) model is a mathematical model that helps predict the probability
that a person will choose a particular travel mode among several available options.
For example, suppose a commuter in Patna must choose between:
Metro,
Government bus, or
Private bus.
Each mode has certain characteristics like cost, travel time, comfort, and safety and each
traveller has their own preferences. The MNL model uses these factors to calculate the
probability that the commuter will select each mode.
Why Do We Use the N-Logit Model in Transportation?
The N-Logit or MNL model is widely used in transportation research for the following reasons:
To understand traveler behavior:
It explains how individuals trade off between different factors, such as saving money
versus saving time.
To forecast future demand:
Transportation planners can use the model to estimate how many people will use a new
metro line or shift to buses if fares are reduced.
To test policy impacts:
The model can evaluate the effects of different transport policies like fare changes,
service improvements, or introduction of new modes (e.g., e-bikes, metro extensions).
To compute the Value of Travel Time (VOT):
The ratio of travel time and cost coefficients helps estimate how much travellers are
willing to pay to save time for cost-benefit analysis.
To support sustainable planning:
By understanding what factors encourage public transport use, cities can design
strategies to reduce car dependency and traffic congestion.
Type of Model
The MNL model belongs to the family of Discrete Choice Models (DCMs).
Discrete Choice Models
Discrete choice models are used when the dependent variable (the outcome) involves a finite
set of alternatives, and the individual can choose only one option at a time.
For example:
Choosing a transport mode (bus, metro, car)
Selecting a route
Choosing a trip timing
Types of Logit Models
There are several logit model types:
1. Binary Logit Model:
Used when there are only two alternatives (e.g., car vs. bus).
2. Multinomial Logit (MNL) Model:
Used when there are more than two alternatives (e.g., metro, government bus, private
bus).
3. Nested Logit Model (N-Logit):
Used when choices are grouped into hierarchies (for example, first choosing between
private and public transport, then choosing between metro and bus within public
transport).
How the Models Works
The basic idea behind the N-Logit (MNL) model is that each person assigns a certain utility
(U) to every available alternative. Utility means the perceived satisfaction or attractiveness of
that alternative to the person. A commuter will choose the mode that provides the highest
utility.
Mathematically, the utility of mode i for person n can be written as:
𝑈 =𝑉 + 𝜀
Where:
𝑈 : Total utility for person n choosing mode i
𝑉 : Systematic or observed component of utility (depends on measurable factors)
𝜀 : Random or unobserved component (accounts for personal preferences or unknown
factors)
The systematic component 𝑉 is generally expressed as a linear function:
𝑉 = 𝛽 ∗ 𝑇𝑇 + 𝛽 ∗ 𝑇𝐶 + 𝛽 ∗ 𝐶𝐹 + 𝛽 ∗ 𝑆𝐹
Where:
𝑇𝑇: Travel time
𝑇𝐶: Travel cost
𝑇𝐶: Comfort
𝑆𝐹: Safety
𝛽 : Coefficients estimated from data (showing sensitivity to each variable)
Probability of Choosing a Mode
The probability that a commuter n chooses mode i is given by the logit formula:
𝑒
𝑃 =
∑ 𝑒
This means:
The higher the utility of a mode, the higher its probability of being chosen.
All probabilities add up to 1 across all modes.
For example, if a traveller has the following utilities:
Metro: 2.5
Government Bus: 1.5
Private Bus: 1.0
Then,
.
P (Metro) = . . .
This gives the probability that the traveller will select the metro.