Reviewed by
Andreas Koch
University of Salzburg
The book Urban Dynamics and Simulation Models presents major results of a project called “GeoDiverCity” which has been conducted by the editors and their colleagues at different scientific institutes, mostly in France. This fact is important because it helps to better understand the intention and the argumentative style of the book.The book is organised in six chapters dealing mainly with the methodology of a specific model type – the agent-based SimpopLocal Model – that is used to investigate the complex dynamics of systems of cities on different geographical scales. Chapter 2 delineates the most relevant characteristics of the SimpopLocal model, while Chapter 3 presents an approach to how to evaluate this model. Chapter 4 discusses a multi-modelling method by incorporating different theoretical approaches to the composition and variation of systems of (interconnected) cities. Chapter 1 introduces the methodological challenges theoretically by asking “is urban future predictable?”. Chapter 5 offers a description of the opportunities to explore the parameter space concerning the possible futures of systems of cities. Finally, chapter 6 briefly introduces the software platform OpenMOLE that has been used to create and run the SimpopLocal models.
The argumentative style is shaped by a strong justification for the chosen approach in the project and the way in which the topic of the dynamics of systems of cities has been investigated. The strength of this style lies in a concise delineation of the context of discovery and the context of justification. Problems of data calibration, model validation and interpretation of results can then be presented convincingly and coherently. The ideas of an “incremental multi-modelling method” (chapter 4) and of “exploring contingency and complexity” (chapter 5) are good examples of this approach.
The weakness of this style, however, is that the reader gets the impression that this approach is the only appropriate one for tackling and grasping the problem. This impression starts with the selection of theories in the book which are limited to rank-size rules, in their different versions, to deal with urban complexity. Even though these rules may exhibit some descriptive power, the explanatory power that the statistical correlation between city size and city rank may have to understand the emergence and consolidation of networks of cities (or city systems) has not been sufficiently outlined. The book lacks hints that there are other methodological approaches as for example network analysis, and other theoretical approaches. No references are presented, for example, to Peter Taylor’s (2003) book about “World City Network. A Global Urban Analysis”, Saskia Sassen’s (2001) book about “The Global City: New York, London, Tokyo”, Manfred M. Fischer’s (2010) approach of spatial interaction modelling (in the book “Spatial Analysis and GeoComputation”), or Manuel Castells’s (1996) work of “The Rise of the Network Society”, which are all seminal contributions to problems and questions raised explicitly in this book.
Related problems, not only given with this book, are the problems of over-generalisation and stating facts as natural facts. I would, for example, hesitate to confirm the statement: “because of our observation that cities never grow in isolation but always through interaction and co-evolution with other cities …” (page xvii) due to its totalitarian inclusion of “never” and “always”. The same doubts are justified against stated facts such as “since systems of cities are complex systems based on nonlinear interactions” (page xvii) while there is no further theoretical grounding of why and how complexity seems to be an appropriate approach in dealing with systems of cities.
Unfortunately, the book also has a couple of formal weaknesses. They start with the title which is somewhat misleading. In fact, the book does not deal with “urban dynamics” but with “systems of cities” and their development which are to be assumed to be dynamic. This assumption has mostly been proven to be true, but due to the framing of the model purposes, these proofs tend to be in the style of self-fulfilling prophecies. I will come back to this issue. Another formal problem is given with the figures: while the legends of the figures in chapter 2 are all in French, figures of chapters 4 and 5 are in bad resolution. Chapter 5.2.1 ends abruptly in the middle of the sentence, and the last complete sentence of this chapter confuses the reader with “… and using Novelty Search to guide the exploration of parameter space in relation to the parameter space”. I would assume that it is the pattern space that is evaluated through an exploration of the parameter space. This assumption would be in line with the title of this chapter (The Pattern Space Exploration Algorithm). The acronym PSE used in this chapter stands for Pattern Search Exploration (page 84) instead of Pattern Space Exploration. All this hampers a reading of the book. The presentation of the chapters is also somewhat odd: the first four follow a coherent structure by presenting the title followed by the text. Chapter 5 differs in that it explicitly mentions the authors of the chapter, followed by keywords after the abstract. And the last chapter of the book has no number and appears to be a bit isolated from the remaining chapters.
Just one word about the problem of self-fulfilling prophecies, which is likewise not restricted to this book. The idea to build models of parsimonious utilisation of variables is tempting. One problem, however, might be the positivistic “you get what you want” philosophy if parsimony equates with over-simplification. Chapter 2 presents an explicitly simple model to analyse the relationships between settlement size, innovation, and allocation of resources. Though simple it is questionable to fully conflate innovation creation with settlement size as has been done here: the model “makes the process of innovation endogenous by linking it with the size of the settlement” (page 22).
To sum up: readers who are already familiar with complexity theory and are interested in relating the development of systems of interconnected cities with rank-size rules or wish to use the open software tool OpenMOLE will find a lot of sophisticated material and further assistance. Those who want to learn more about urban development are pointed to the references presented in the book.
ReferencesFISCHER Manfred M. (2006). Spatial Analysis and GeoComputation. Springer.
SASSEN Saskia (2001). The Global City: New York, London, Tokyo. Princeton Paperbacks.
TAYLOR Peter (2003). World City Network: A Global Urban Analysis. Routledge.
Return to Contents
of this issue