Repository logo
 

Refining Behavioural Theories and Rules in Agent-Based Models to Enhance Dynamic Simulation of Urban Change


Change log

Authors

Abstract

With the continuing growth of the urban population, many cities in the world are faced with challenges such as traffic congestion and misallocation of housing and infrastructure. Understanding and predicting the spatial pattern of urban change is useful for planners to deliver evidence-based and adaptive policies for addressing current problems and pursuing future sustainability. Harnessing technologies such as geographic information system (GIS), approaches like agent-based modelling (ABM) have been applied to generate dynamic simulations to analyse changes in the urban landscape. However, the current approach to ABM tends to predict behaviours based on a few quantitative factors rather than taking various behavioural theories and rules into sufficient account. There is a need to better understand the ranges of behaviours and theories, and the associated patterns and processes. Also, the current approach that relies on the mathematical realm of research should expand and include the psychological and sociological realms.

To address these gaps in the literature, this PhD research first maps sixty-two behavioural theories in a diagram and discusses the role of behavioural theories in bridging the theory-driven and data-driven approaches to research. Then, it zooms into planning-related fields and proposes guidance for linking behavioural theories with types of behaviour, key variables, rules, and research methods, and discusses its applicability in urban models such as space and time-sensitive dynamic simulation. Based on this theoretical understanding, an empirical research design gets established to demonstrate how behavioural theories and rules can be refined in an ABM to simulate the dynamic interaction between human behaviour and urban space. Using data from Sejong, Korea, the thesis explains the various decisions made during the technical process of preparing spatial data and extending the existing CA-based SLEUTH urban growth model into a CA/ABM-based land use-transport interaction model in the NetLogo platform. Following on, it demonstrates the structuring of behavioural rules for residents’ car to non-car mode switch based on three psychosocial theories using real spatial data and empirical behavioural data of agent characteristics. Furthermore, it provides policy implications based on the simulation results in consideration of the aspect of value judgement and emphasises the importance of applying both top-down and bottom-up approaches for encouraging behaviour change.

Description

Date

2020-11-20

Advisors

Silva, Elisabete

Keywords

agent-based modelling, urban and environmental planning, complexity theory, behavioural theories and rules, behavioural sciences, dynamic simulation, space and time interaction models, land use-transport interaction model, SLEUTH urban growth model, cellular automata

Qualification

Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge