Doctoral thesis (Dissertations and theses)
An agent-based framework for modeling travel behavior under disrupted networks
Saadi, Ismaïl
2017
 

Files


Full Text
final.pdf
Author preprint (597.59 kB)
The full manuscript will be online soon
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Abstract :
[en] Worldwide, floods are the most frequent natural disasters and cause over one third of overall economic losses due to natural hazards. In addition, studies show that flood risk will further increase during the 21st century as a result of climate change. In this context, the objective of this thesis is to implement a methodological framework for modeling the impact of disrupted networks caused by river floods on travel demand. The manuscript is divided into four chapters. In the first chapter, this thesis contributes to the state-of-the-art by presenting an integrated multidisciplinary framework capable of making long-term projections (time horizon 2050 and 2100) with the objective of mitigating future flood risk. Various aspects of modeling are discussed with a focus on the interactions between the different model components. The second chapter is dedicated to multi-source data fusion in transportation research. Indeed, conducting large scale data collection is difficult and requires substantial financial resources. In practice, micro-samples with small sampling rates are generally used for synthesizing populations of households and individuals. Unfortunately, they present important limitations from a qualitative point of view, i.e. lack of representativeness. In this regard, a full population synthesis procedure based on Hidden Markov Model (HMM) has been designed to enable multi-source data fusion and incorporate more heterogeneity into eventual poor data-sets. Our research revealed that HMM outperforms IPF for all the sampling rates smaller than 25% regardless the scalability, while the amount of input data used by HMM is lower compared with IPF. The characterization of activity-travel patterns is described in the third chapter. Indeed, to enable a better understanding of travel behavior, a simulation-based approach for population synthesis has been coupled with a profile Hidden Markov Model (pHMM) in laying the foundation for an innovative activity-based model. There have been several key issues that deserved special attention, in particular the influence of socio-demographics on the activity-travel patterns. We also proposed new perspectives of validation techniques. The simulated and observed activity-travel patterns have been systematically compared on the basis of the emission and transition probabilities. The fourth chapter is dedicated to the case study. The synthesized population is used to calibrate a large scale scenario based on the agent-based framework MATSim to investigate the short-term impact of river floods on travel demand. The model has been tested for Liège, Belgium with multiple flood risk scenarios. Results reveal that the impact in terms of travel times is particularly significant when the network is operating at capacities lower than 50%.
Research center :
Lepur : Centre de Recherche sur la Ville, le Territoire et le Milieu rural - ULiège
UEE - Urban and Environmental Engineering - ULiège
Local Environment & Management Analysis (LEMA)
Disciplines :
Engineering, computing & technology: Multidisciplinary, general & others
Author, co-author :
Saadi, Ismaïl ;  Université de Liège - ULiège > Département ArGEnCo > Transports et mobilité
Language :
English
Title :
An agent-based framework for modeling travel behavior under disrupted networks
Defense date :
November 2017
Institution :
ULiège - Université de Liège
Degree :
Doctorat en sciences de l'ingénieur
Promotor :
Cools, Mario  ;  Université de Liège - ULiège > Urban and Environmental Engineering
Teller, Jacques  ;  Université de Liège - ULiège > Département ArGEnCo > Urbanisme et aménagement du territoire
President :
Limbourg, Sabine  ;  Université de Liège - ULiège > HEC Recherche > HEC Recherche: Business Analytics & Supply Chain Management
Jury member :
Dewals, Benjamin  ;  Université de Liège - ULiège > Urban and Environmental Engineering
Farooq, Bilal
Witlox, Frank
Name of the research project :
ARC - Floodland
Funders :
ARC grant for Concerted Research Actions financed by the Wallonia-Brussels Federation
Available on ORBi :
since 25 November 2017

Statistics


Number of views
218 (53 by ULiège)
Number of downloads
17 (15 by ULiège)

Bibliography


Similar publications



Contact ORBi