Optimal Transport in Worldwide Metro Networks

TL;DR

This paper analyzes 28 major metro networks worldwide using optimal transport metrics, revealing geometric and transport property relationships, and ranking networks based on key transport measures.

Contribution

It introduces a comparative analysis of metro networks using Wasserstein distance and geometric measures, highlighting relationships and network rankings.

Findings

Higher fractal dimension correlates with lower energy costs.

Wasserstein distance relates to fractal dimension and transfers.

New York and Berlin metros rank highest in transport efficiency.

Abstract

Metro networks serve as good examples of traffic systems for understanding the relations between geometric structures and transport properties.We study and compare 28 world major metro networks in terms of the Wasserstein distance, the key metric for optimal transport, and measures geometry related, e.g. fractal dimension, graph energy and graph spectral distance. The finding of power-law relationships between rescaled graph energy and fractal dimension for both unweighted and weighted metro networks indicates the energy costs per unit area are lower for higher dimensioned metros. In L space, the mean Wasserstein distance between any pair of connected stations is proportional to the fractal dimension, which is in the vicinity of our theoretical calculations treated on special regular tree graphs. This finding reveals the geometry of metro networks and tree graphs are in close proximity to one another. In P space, the mean Wasserstein distance between any pair of stations relates closely to the average number of transfers. By ranking several key quantities transport concerned, we obtain several ranking lists in which New York metro and Berlin metro consistently top the first two spots.