Roads patterns following biological patterns
Belle Dumé in the NewScientist addressed recently the idea that city road networks grow like biological systems. The article is basically a description of the academic work of Marc Barthélemy and Alessandro Flammini who analysed street pattern data from roughly 300 cities, including Brasilia, Cairo, Los Angeles, London, New Delhi, and Venice.
Using these cases, the researchers found interesting patterns showing that the road networks in cities evolve driven by a simple universal mechanism that follows a biological metaphor:
“The main influence on the simulated network as it grows is the need to efficiently connect new areas to the existing road network – a process they call “local optimisation”. They say the road patterns in cities evolve thanks to similar local efforts, as people try to connect houses, businesses and other infrastructures to existing roads.
(…)
“Beyond the economic, demographic and geographic “forces” that shape a town, there are a myriad of small “accidents” that contribute” he says. “Although these are unpredictable, they can be understood in terms of statistics and simple modelling.”The team’s model also reveals that roads often bend, even in the absence of geographical obstacles, and that road intersections are generally perpendicular.“
And, as the authors described in their paper, “in the absence of a global design strategy, the evolution of many different transportation networks indeed follows a simple universal mechanism.”
Why do I blog this? I am not really into urban pattern modeling but I find interesting this notion of “local optimisation” and how it works for instance for roads and not for rail (because of its different nature and scales).
This is somewhat related to the elephant path (desire line) I often blog about here and there as pointed out by Space and Culture. A desire line can be turned into a design opportunity and thus into a new road.
Why is that interesting? certainly because it shows the contingencies of the urban infrastructure. I am wondering this hold true for other sort of infra, such as internet connections.
April 28th, 2008 at 1:05 am
This is a fascinating post, and yet I wonder as to the degree of difference between Dumé’s findings and those of Walter Christaller’s Die zentralen Orten im Süddeutschland (1933) (translated as Central Places in Southern Germany). In this text, Christaller articulated a system of relationships between urban centers and the surrounding lands. These relationships are both qualitative and quantitative — qualitative in the sense that Christaller himself uses the term “optimization of the distribution of cities”; qualitatitve, however, in the sense that he uses a system of hexagons to represent this idea of optimization. The real conceptual difference, however, is that Christaller’s Central Place theory is first and foremost a model, and one that occurs in a geographically-impossible flat, isotropic plane. His was an influential theory, laying the groundwork for C.A. Doxiadis’ work on Classical Greek infrastructural planning, to William Cronon’s magisterial Nature’s Metropolis: Chicago and the Great West (1992).
I also note that using biological examples to explain the particularities of a system has a very distinguished pedigree. But more recently, German sociologist Niklas Luhmann used the theories of Chilean biologists Humberto Maturana and Francisco Varela for his thrilling and maddening Soziale Systeme (1984) (translated as Social Systems in 1995). Though Luhmann probably never conceived of his system as a way to explain infrastructural phenomena, you can imaging that some enterprising student of regional and urban planning will.