Please use this identifier to cite or link to this item:
|Title:||Modelling heterogeneous decision processes and joint decision-making in travel demand models||Authors:||Blomberg Stathopoulos, Amanda Irini||Supervisore/Tutore:||Marcucci, Edoardo||Cosupervisore:||Danielis, Romeo||Issue Date:||19-Apr-2012||Publisher:||Università degli studi di Trieste||Abstract:||
There is substantial interest in encouraging changes to travel behaviour with a view to accomplishing more sustainable mobility patterns.
The underlying idea is that people respond to incentives and will alter their behaviour according to relative costs and benefits of different behavioural alternatives (e.g. the use of different transport modes for the commute trip). Utility-based discrete-choice models have become central methods to model behaviour with the aim of understanding how changes can be induced. Traditionally such models, however, assume that choices can be represented as a linear compensatory process. This implies that there is trading among attributes, that is, disadvantages in one choice characteristic can be offset by advantages in another. Similarly standard modelling assumptions postulate that group behaviour can be represented through a one-consumer utility function. This implies that the study of essential economic group-based agents take account of only a single representative of the entity, without considering the impact of the presence of different members.
Applied studies of real behaviour has generated many findings suggesting that people use non-maximising rules and that multi-person choices are different from individual ones. Failing to account for decisions that do not adhere to these underlying hypotheses, may generate biased descriptions and predictions of behaviour. A poor understanding of real behavioural motivations will potentially lead to misguided policy decisions. This thesis proposes the study of several failures of standard modelling assumptions. Methodological approaches, where standard modelling procedures are adapted, are illustrated in this work. The advantage of the proposed approaches is to gain a deeper understanding of behaviour and begin staking out how people differ not only in their taste structure but also along other behavioural dimensions.
Evidence from four empirical studies are presented. A first case-study looks at the role of reference dependence, focussing on multiple attributes and multiple reference points in a commuting context. This allows in depth study of the usual assumption of reference free, linear and symmetrical sensitivities. The second chapter looks at a modelling structure that can account for different decision-rules, besides utility-maximisation, that can be used to model decision-rules such as lexicography, reference asymmetry, elimination-by-aspects and regret minimisation. This approach offers a way to relax the assumption that all respondents use utility maximising decision protocols. A third section describes a model structure where the level of engagement of respondents is studied using a latent variable structure to see how involvement can be studied from attitudinal questions and other behavioural variables. This is a way to assess the impact of lower involvement in a survey leading to higher error variance in responses rather than assuming all respondents to be equally engaged in experimental tasks. A fourth chapter overlooks a framework of individual versus joint preference formation in a household to understand the potential shortcomings of the representative respondent hypothesis.
The results show how different behavioural model assumptions can be tested within a discrete choice framework. Each case shows that modelling can be improved upon by allowing people to differ in referencing, behavioural rules, survey engagement and in a joint choice context. Taken together, these findings help us bridge the gap between observed behavioural complexity and the use of formal models of decision-making.
|Ciclo di dottorato:||XXIV Ciclo||metadata.dc.subject.classification:||SCUOLA DI DOTTORATO DI RICERCA IN SCIENZE INTEGRATE PER LA SOSTENIBILITA' TERRITORIALE||Description:||
|Keywords:||discrete choice; decision process; joint choice||Type:||Doctoral||Language:||en||Settore scientifico-disciplinare:||SECS-P/06 ECONOMIA APPLICATA||NBN:||urn:nbn:it:units-9183|
|Appears in Collections:||Scienze economiche e statistiche|
Show full item record
Files in This Item:
Page view(s) 201,060
checked on Oct 27, 2020
checked on Oct 27, 2020
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.