Agent-based modelling of people movement using mixed methods approach
File(s)
Author(s)
Cheong, He-in
Type
Thesis
Abstract
Existing agent-based modelling methods of people movement suffer from numerous drawbacks, such as their assumptions having little reference to any evidence, their lack of consideration of qualitative data from the research subjects, and ethical issues that may arise from current model development processes. Therefore, this thesis considers different fields of people movement of building evacuation, road transport, future scenarios of autonomous vehicles, and forced migration, and establishes a sequential mixed methods approach based on the philosophy of pragmatism to develop an agent-based modelling framework that explicitly considers ethics, qualitative as well as quantitative data, and validation methods evaluating the internal validity of the models. As part of this, a novel methodology for secondary qualitative research is developed that focuses on quality assessment to mitigate bias whilst upholding research ethics. The proposed mixed methods approach is applied to the case study of forced migration through triangulation of qualitative research, quantitative research, and agent-based modelling, whereby each module of the research process is clearly linked with each other through information transfer and development of new research questions at each stage. The scope of the thesis covers a spatiotemporally diverse geographical study area over three continents for a time period between 2010 and 2019 via a dataset of 36 transcripts post quality assessment of over 200 transcripts and 30 types of quantitative data of multiple sources. The results from the quantitative and qualitative research define the properties and key performance indicators of different agent types, topology, and environment and specify over 20 parameters and the cognitive and behavioural logic for the forced migrant. The assessment shows that the developed model agrees with research outputs of other qualitative researchers and is far more complex than existing agent-based models of forced migration and explicitly addresses ethics at each stage of the modelling process.
Version
Open Access
Date Issued
2023-09-08
Date Awarded
2024-03-01
Copyright Statement
Creative Commons Attribution NonCommercial NoDerivatives Licence
Advisor
Majumdar, Arnab
Ochieng, Washington
Sponsor
Lloyd’s Register Foundation
Engineering and Physical Sciences Research Council
Imperial College London
Grant Number
EP/L016826/1
Publisher Department
Civil and Environmental Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)