Agent-based Models for Evaluating Sustainable Transportation Systems
Doctoral thesis, 2024
Many countries and organisations are adopting strategies to promote a sustainable transition of the transportation system and mitigate the adverse effects of climate change. To guide these efforts, decision-makers need tools that are capable of evaluating complex transportation policies and assessing sustainable transportation scenarios.
This thesis presents the development of transportation models that simulate individuals' travel behaviours and evaluate sustainable transportation technologies. Agent-based models (ABMs) are well suited to address the complexities of transportation systems through individual-level modelling, while activity-based travel demand generation integrates the behavioural dynamics driving demand into the model. Specifically, the thesis aims to answer three research questions: (1) How can a synthetic population of Sweden be created for use in ABMs? (2) How can realistic travel demand be generated for the developed population? (3) How can state-of-the-art methods and big data sources be incorporated into the modelling process?
Papers I and II introduce the Synthetic Sweden Mobility (SySMo) model, a large-scale agent-based transportation model. Paper I details the creation of a synthetic population for the entire Swedish population, addressing a significant data gap (Q1). Paper II focuses on SySMo's travel demand generation module, proposing a novel methodology that preserves heterogeneity in travel demand using machine learning algorithms (Q2-Q3). Paper III explores the integration of big data sources into activity-based models, introducing a generative model that synthesises multi-day activity-travel plans for over 263,000 individuals residing in Sweden using mobile data (Q3). Paper IV demonstrates an application of SySMo by analysing sustainable travel modes, specifically assessing the potential of e-bikes to replace passenger car trips and reduce greenhouse gas emissions.
The developed models may assist decision-makers by capturing individuals' travel behaviours and evaluating emerging technologies, enabling informed policy formulation. To support collective efforts toward sustainable transportation, this research adheres to open science principles by making the models publicly accessible.
This thesis presents the development of transportation models that simulate individuals' travel behaviours and evaluate sustainable transportation technologies. Agent-based models (ABMs) are well suited to address the complexities of transportation systems through individual-level modelling, while activity-based travel demand generation integrates the behavioural dynamics driving demand into the model. Specifically, the thesis aims to answer three research questions: (1) How can a synthetic population of Sweden be created for use in ABMs? (2) How can realistic travel demand be generated for the developed population? (3) How can state-of-the-art methods and big data sources be incorporated into the modelling process?
Papers I and II introduce the Synthetic Sweden Mobility (SySMo) model, a large-scale agent-based transportation model. Paper I details the creation of a synthetic population for the entire Swedish population, addressing a significant data gap (Q1). Paper II focuses on SySMo's travel demand generation module, proposing a novel methodology that preserves heterogeneity in travel demand using machine learning algorithms (Q2-Q3). Paper III explores the integration of big data sources into activity-based models, introducing a generative model that synthesises multi-day activity-travel plans for over 263,000 individuals residing in Sweden using mobile data (Q3). Paper IV demonstrates an application of SySMo by analysing sustainable travel modes, specifically assessing the potential of e-bikes to replace passenger car trips and reduce greenhouse gas emissions.
The developed models may assist decision-makers by capturing individuals' travel behaviours and evaluating emerging technologies, enabling informed policy formulation. To support collective efforts toward sustainable transportation, this research adheres to open science principles by making the models publicly accessible.
Activity-based modelling
Synthetic population
Activity-travel plans
Big data.
Agent-based modelling
Room EC, EDIT huset, Hörsalsvägen 11, Gothenburg
Opponent: Prof. Dr. Yusak Susilo, University of Natural Resources and Life Sciences (BOKU), Vienna, Austria
Author
Çaglar Tozluoglu
Chalmers, Space, Earth and Environment, Physical Resource Theory
Have you ever thought how the daily mobility patterns of millions could be captured to evaluate strategies supporting the sustainable transition of transportation systems?
This thesis presents the development of decision-support tools that capture individual travel behaviours and evaluate sustainable transportation scenarios in Sweden. It introduces the Synthetic Sweden Mobility (SySMo) model, a large-scale agent-based transportation model. SySMo generates a detailed synthetic population, incorporating various socio-demographic variables and activity-travel schedules across the entire Swedish population, using state-of-the-art methods from transportation modelling and computer science. Additionally, this thesis explores integrating big data sources into activity-based travel demand generation, introducing a novel model that synthesises activity-travel plans from extensive but incomplete mobile phone application data. It also demonstrates an application of SySMo by assessing the potential of e-bikes to replace passenger car trips and reduce greenhouse gas emissions.
These models aim to assist decision-makers by capturing individuals' travel behaviours and evaluating emerging technologies, enabling informed policy formulation. To support collective efforts toward sustainable transportation, this research adheres to open science principles by making the models publicly accessible.
This thesis presents the development of decision-support tools that capture individual travel behaviours and evaluate sustainable transportation scenarios in Sweden. It introduces the Synthetic Sweden Mobility (SySMo) model, a large-scale agent-based transportation model. SySMo generates a detailed synthetic population, incorporating various socio-demographic variables and activity-travel schedules across the entire Swedish population, using state-of-the-art methods from transportation modelling and computer science. Additionally, this thesis explores integrating big data sources into activity-based travel demand generation, introducing a novel model that synthesises activity-travel plans from extensive but incomplete mobile phone application data. It also demonstrates an application of SySMo by assessing the potential of e-bikes to replace passenger car trips and reduce greenhouse gas emissions.
These models aim to assist decision-makers by capturing individuals' travel behaviours and evaluating emerging technologies, enabling informed policy formulation. To support collective efforts toward sustainable transportation, this research adheres to open science principles by making the models publicly accessible.
Subject Categories
Other Computer and Information Science
Transport Systems and Logistics
ISBN
978-91-8103-135-5
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5593
Publisher
Chalmers
Room EC, EDIT huset, Hörsalsvägen 11, Gothenburg
Opponent: Prof. Dr. Yusak Susilo, University of Natural Resources and Life Sciences (BOKU), Vienna, Austria