Structure Entropy, Self-Organization, and Power Laws in Urban Street Networks: Evidence for Alexander's Ideas

Abstract

Easy and intuitive navigability is of central importance in cities. The actual scale-free networking of urban street networks in their topological space, where navigation information is encoded by mapping roads to nodes and junctions to links between nodes, has still no simple explanation. Emphasizing the road-junction hierarchy in a holistic and systematic way leads us to envisage urban street networks as evolving social systems subject to a Boltzmann-mesoscopic entropy conservation. This conservation, which we may interpret in terms of surprisal, ensures the passage from the road-junction hierarchy to a scale-free coherence. To wit, we recover the actual scale-free probability distribution for natural roads in self-organized cities. We obtain this passage by invoking Jaynes's Maximum Entropy principle (statistical physics), while we capitalize on modern ideas of quantification (information physics) and well known results on structuration (lattice theory) to measure the information network entropy. The emerging paradigm, which applies to systems with more intricate hierarchies as actual cities, appears to reflect well the influential ideas on cities of the urbanist Christopher Alexander.