Generative network modelling: a comparative study on rail and road networks in Yangtze River Delta

This paper focuses on the modelling and simulation of the inter-city transport connections underlying urban networks. Inspired by recent advances that aim to bridge spatial science with network science, our study draws on Vértes et al.'s (2012) economical clustering model and proposes a generative network model (GNM) to integrate factors commonly captured in traditional spatial simulation models (e.g., gravity models) and topological simulation models (e.g., actor-oriented stochastic models) into a single framework. In our transportation network-implementation of GNM, it is assumed that the emergence of inter-city linkages can be approximated through probabilistic processes that speak to a series of contradictory forces. In our exploratory study, hampering factors include distance and administrative boundaries, while stimulating factors include city size and a topological rule stating that there is more likelihood in the formation of connections between cities sharing nearest neighbours (i.e., a transitive effect). Based on a comparative study of the modelling of rail and road networks in the highly urbanized Yangtze River Delta (China), two practical implications of this approach are discussed: (1) it confirms the potential of the proposed method for urban network simulation, as the inclusion of a topological factor alongside spatial factors generates a network that better approximates the observed one; and (2) we can explore the different extent to which driving forces shape the different transport networks. For instance, the transitive effect plays an equally important role in the formation of both transport networks in the study area; GDP has a bigger impact than population in the generation of the rail network, while the obverse holds in the road network; the boundary effect plays a more prominent role in shaping the rail network; and the distance decay effect is stronger in the road network