Shape and Performance of Fastest Paths over Networks with Interacting Selfish Agents

Marco Cogoni,Giovanni Busonera,Enrico Gobbetti

2024-12-23

Abstract:We study the evolution of the fastest paths in transportation networks under increasing congestion, modeled as a linear decrease in edge travel speed with density due to interactions among selfish agents. Moving from the common edge-based to a path-based analysis, we examine the fastest directed routes connecting random origin-destination pairs as traffic grows, characterizing their shape through effective length, maximum detour, and area under the curve, and their performance through a novel metric measuring how fast and how far an agent travels toward its destination. The entire network is characterized by analyzing the performance metric's distribution across uniformly distributed paths. The study covers both random planar networks with controlled characteristics and real urban networks of major cities. The low-density network regime, in which an initial smooth performance degradation is observed up to a critical traffic volume, is followed by the emergence of complex patterns of spatially heterogeneous slowdowns as traffic increases, rapidly leading to disjoint subnetworks. The failure of a few edges leads to a catastrophic decrease in the network performance. The fastest paths for all cities show a peak for detour and inness (and their variance) in the proximity of the critical traffic level, defined as the flex of the rejected path ratio curve. Inness generally shows a slight attraction by city centers on paths for light traffic, but this reverses to strong repulsion during congestion. We exploit path performance to uncover an asymmetric behavior of different regions of the networks when acting as origins or destinations. Finally, the Gini coefficient is used to study the unequal effects of path performance degradation with traffic.

Physics and Society,Statistical Mechanics