In transportation modeling, both the trip interchange model split and the
aggregate/disaggregate model split are important approaches to modeling
and forecasting travel behavior and patterns. These splits refer to how trips
or travel behaviors are categorized and analyzed in models, particularly in
transportation planning and traffic management.
1. Trip Interchange Model Split
A trip interchange model split focuses on the distribution and categorization
of trips between origins and destinations (OD pairs). It is primarily used in
trip distribution models, which aim to understand how many trips occur
between different zones or areas within a region.
Key Components:
Trip Distribution: Determines the number of trips between different origin-
destination pairs (zones), typically based on factors like distance, time, and
the attractiveness of the destination.
Mode Choice: Once trips are distributed, the model determines the mode of
transportation (car, bus, train, bike, walking, etc.) for each trip. This is the
“mode split” aspect of the trip interchange model.
Purpose of the Trip: Trips are often categorized based on purpose
(commuting, shopping, recreational) since the trip purpose influences mode
choice and travel patterns.
Split Methodology:
Friction Factors: The model assigns “friction” to travel based on time or
distance. Longer trips have more friction, reducing the likelihood of travel.
Gravity Model: A common approach in trip interchange models, which states
that the number of trips between two zones is proportional to the population
(or economic activity) of those zones and inversely proportional to the
distance or travel time between them.
Examples:
Urban Transport Planning: A city could use a trip interchange model split to
understand how many people will commute between the suburbs and the
central business district (CBD), and then split that into different transport
modes (e.g., car, bus, subway).
Intercity Travel: The model can predict how many people will travel between
two cities and the mode they will likely choose based on distance, time, and
availability of transport modes (e.g., high-speed rail vs. driving).
Challenges:
Complexity in Mode Split: As transportation networks become more complex
(with multimodal options like ridesharing, micromobility, etc.), accurately
splitting trips by mode becomes more difficult.
Behavioral Variations: Different trip purposes and individual preferences can
cause significant variations in travel behavior that need to be modeled
accurately.
2. Aggregate and Disaggregate Model Split
The aggregate and disaggregate model split refers to how travel behavior or
trip data is analyzed at different levels of detail—whether at the individual
level (disaggregate) or at a group or zonal level (aggregate).
1. Aggregate Model Split
Definition: In an aggregate model split, the model looks at travel patterns for
groups or categories of travelers, typically grouped by geographic zones,
demographic characteristics, or trip purpose.
Focus: The model outputs data that represents the average behavior or
overall trends for a group of individuals rather than focusing on individual
travel decisions.
Example: A transportation model that predicts how many people in a specific
neighborhood will use public transportation as opposed to driving, without
differentiating between individual commuters.
Mode Split: The mode choice for trips is often calculated as a proportion of
the total trips. For example, 60% of all trips in a region may be by car, 30%
by public transit, and 10% by walking/biking.
Pros of Aggregate Model Split:
Simplicity: Aggregated models are simpler and computationally less
expensive.
Effective for Broad Forecasting: Useful for overall transportation planning
where high-level patterns are sufficient.
Cons of Aggregate Model Split:
Lack of Granularity: It doesn’t capture individual travel behaviors and
variations in preferences.
Homogenization: Aggregate models assume homogeneity within groups,
which might not reflect real-world travel behaviors.
2. Disaggregate Model Split
Definition: In a disaggregate model split, individual travel behavior is
modeled, allowing for more detailed insights into how and why specific
people make specific travel decisions.
Focus: Disaggregate models focus on the choices made by individuals or
households, using data from travel surveys, demographics, and specific
travel patterns.
Example: A model that predicts whether a specific person with certain
socioeconomic characteristics (e.g., income level, household size) will use a
bike, car, or bus to travel to work.
Mode Split: Disaggregate models determine mode choice at the individual
level, considering personal characteristics like income, age, and preferences.
These models often use utility-based approaches, such as **discrete choice
models**, to predict the probability that an individual will choose a specific
mode.
Pros of Disaggregate Model Split:
Detailed Insights: Captures variations in individual behavior, which is
particularly useful for understanding diverse travel patterns.
Behaviorally Realistic: Can incorporate individual preferences, constraints,
and perceptions.
Cons of Disaggregate Model Split:
Data-Intensive: Requires detailed data on individuals, which can be difficult
and expensive to collect.
Computationally Complex: Disaggregate models require more computational
power and resources to run simulations at an individual level.
Engineering the Model Splits:
Trip Interchange Model Engineering
Inputs: The inputs include origin-destination pairs, population data, trip
purpose data, and transportation networks.
Mode Split Component: A mode choice model is applied to distribute the trips
into different modes of transport.
Calibration: The model is calibrated using historical data, survey data, or
travel diaries to predict future trip patterns and mode choices.
Aggregate/Disaggregate Model Engineering:
Aggregate Model Engineering: Typically involves creating zonal aggregates,
grouping trips by region or demographic, and using statistical techniques to
model the mode split for the entire group. Aggregated data sources such as
census data or large-scale surveys are often used.
Disaggregate Model Engineering: Uses individual-level data to build utility-
based models, like logit models, which can predict individual travel behavior
and mode choice by maximizing the perceived utility for each option.
Requires advanced data collection (e.g., household travel surveys).
Conclusion:
Trip interchange model split is used to distribute trips between regions or
zones and split them by mode, focusing on where trips are made and how
they are carried out across zones.
Aggregate and disaggregate model split deals with how data is grouped or
individualized in transportation models, with aggregate models providing a
broad overview and disaggregate models providing detailed, person-level
insights.
Both methodologies are valuable depending on the transportation planning
context and the level of detail required.
