Modeling of Transport Demand

Profillidis and Botzoris have written a welcome addition to what is only a handful of key textbooks on a topic that remains largely misunderstood - to develop trustworthy, transparent, and ultimately useful data driven, mathematical tools to support decision making in a highly politicized transport field that also seems to be in greater flux than we have experienced in the last few decades.
This is not a book about using transport models for policy development, or as input to cost benefit analysis. Other books cover that well, for example Ortuzar and Willumsen’s 2011 book “Modelling Transport”. The main goal of this book is to support commercial companies and government entities when planning future operational requirements; paraphrasing the back cover: “to help readers solve transport demand forecasting problems they face on a daily basis, choosing the most suitable solution for all types of applications”.
Given the pedigree of the authors, I expect the book to originate from course lectures, and I can imagine the book being used as a textbook in undergraduate or postgraduate courses elsewhere. I would gladly employ students that have been taught and master the contents. I can also imagine the book on the bookshelf of researchers, professionals and practitioners in government organizations and in industry.
The first two chapters provide essential background to the way in which robust transport demand models should be developed, the drivers of demand and the way in which to incorporate these. In a second edition, I hope that the authors will debate how some of the conventional relationships, for example between GDP growth and the demand for travel, may be breaking, and how to estimate robust models in such circumstances.
Chapter 3 (methods of modelling transport demand) should be compulsory reading for all entrants to the profession for whom this would provide an excellent foundation for what is often experienced in everyday work life as just the following of best practice guidelines and the application of commercial off-the-shelf software.
Given that there is such a strong focus in applied transport demand modelling on quantitative methods, the sections in Chapter 4 on Delphi methods and scenario writing are relevant and interesting - particularly in the present climate of uncertainty when quantification may well be impossible and often unrealistic. I am not sure of section 4.4 on survey methods, and better books exist, such as the still excellent 1995 Richardson, Ampt and Meyburg book “Survey methods for transport planning”, and the more recent 2013 book by Zmud, Lee Gosselin, Munizaga and Carrasco “Transport Survey Methods: Best Practice for Decision Making”.
The real strength of the book is in sections 5, 6 and 7. I have not come across such an in-depth description in a single location of how to conduct well statistical, time series and econometric modelling and forecasting studies. Chapter 5 provides an easy-to-access statistical introduction, but also introduces important concepts such as multicollinearity, residuals, heteroskedacity and suitable tests for these – often ignored in practice. The practical value of trend projection and time series methods, as discussed in Chapter 6, may feel somewhat dubious given that various past trends in transport demand appear to be breaking. Still, there will be applications that rely on projections and the examples should ensure that these are carried out in line with good practice.
The real value of chapter 7 is in the development and description of examples of transport demand models based on assumed causality rather than correlation – topical and sometimes contentious. My greatest criticism is the lack of discussion about equilibrium modelling, standard in the majority of practical urban applications.
I like chapters 8 and 9 a lot. These describe Artificial Intelligence, Neural Networks and Fuzzy Methods. The authors manage to engage the reader with what are conceptually complex methods, and give interesting and easy-to-understand examples how these new methods can be used to improve practice.
In conclusion, Profillidis and Botzoris have done an excellent job, creating an accessible text that should be viewed as additional to rather than replacing its better-known predecessors. The modelling of transport demand is a vast area to cover, and much of the existing literature has focused on urban, policy-supporting approaches. Those with a need to also consider the need for application in the transport industry now have access to a single source volume; available as paperback and as e-book.