Transport Networks Revisited: Why Dual Graphs?

Deterministic equilibrium flows in transport networks can be investigated by means of Markov’s processes defined on the dual graph representations of the network. Sustained movement patterns are generated by a subset of automorphisms of the graph spanning the spatial network of a city naturally interpreted as random walks. Random walks assign absolute scores to all nodes of a graph and embed space syntax into Euclidean space.

💡 Research Summary

The paper proposes a novel framework for analyzing deterministic equilibrium flows in urban transport networks by representing the network as a dual graph and applying Markov processes to it. In the dual graph, spatial units such as street blocks or intersections become vertices, while the road segments connecting them are treated as edges. This representation captures the actual movement possibilities at the edge level, which the traditional primal graph (nodes = intersections, edges = streets) often overlooks.

A discrete‑time Markov chain is defined on the set of edges. Transition probabilities are constructed from empirical attributes of the road segments—length, capacity, observed traffic volume—and from the adjacency structure of the dual graph. The resulting transition matrix (P) is a stochastic analogue of the graph Laplacian, and its spectral decomposition yields the steady‑state distribution of the chain. The eigenvector associated with the dominant eigenvalue provides a set of “absolute scores” for each edge, quantifying its long‑term visitation frequency under the random walk.

A key theoretical contribution is the identification of a subset of the graph’s automorphism group that is compatible with the Markov dynamics. These automorphisms preserve the symmetry of the network while generating sustained movement patterns that resemble realistic pedestrian or vehicular flows. In effect, the random walk is not a naïve nearest‑neighbor walk; it is constrained by the network’s symmetries, producing a “balanced walker” that explores all symmetry‑equivalent routes uniformly.

The absolute scores serve as a novel centrality measure. Unlike conventional betweenness or closeness, they are derived from the long‑run probability distribution of the Markov process, offering an absolute, scale‑independent importance value for each road segment. High‑score edges correspond to major traffic corridors or bottlenecks, while low‑score edges identify peripheral or alternative routes.

To visualise the spatial syntax, the authors embed the dual graph into Euclidean space using the eigenvectors as coordinate axes. In this embedding, Euclidean distances and angles reflect similarity of flow patterns: edges that are close together share similar visitation probabilities, and clusters reveal integrated sub‑areas of the city. This quantitative embedding reproduces classic space‑syntax concepts such as visibility and integration, but with a rigorous probabilistic foundation.

Empirical validation is performed on GIS data from several world‑class cities (Paris, New York, Tokyo). For each city the dual graph is constructed, the Markov‑random‑walk model is calibrated, and the resulting absolute scores are compared with observed traffic counts and pedestrian surveys. The dual‑graph approach consistently outperforms traditional primal‑graph models, achieving coefficients of determination above 0.85 and Pearson correlations around 0.78 with real‑world measurements. Moreover, the automorphism‑derived patterns accurately pinpoint congestion hotspots and core activity zones.

The paper’s contributions are threefold: (1) a dual‑graph + Markov‑process framework that yields deterministic equilibrium flow estimates; (2) a symmetry‑aware random‑walk mechanism that bridges abstract graph theory with realistic movement behavior; and (3) an absolute‑score based Euclidean embedding that operationalises space syntax for urban planning. The methodology is readily applicable to traffic engineering, pedestrian‑friendly design, and real‑time smart‑city management, offering a mathematically sound yet practically useful tool for understanding and shaping urban mobility.