An Event–Link Network Model Based on Representation in P-Space

TL;DR

This paper introduces a new network model based on event–link relationships in P-space, which can generate networks with properties similar to real-world systems like transportation networks.

Contribution

The novel event–link model in P-space enables the generation of networks with controllable topological features and realistic statistical properties.

Findings

The event–link model generates networks with small-world and scale-free properties.

Simulation outcomes align with theoretical predictions under various parameter settings.

The model can replicate the growth and structure of real-world transportation networks.

Abstract

The L-space and P-space are two essential representations for studying complex networks that contain different clusters. Existing network models can successfully generate networks in L-space, but generating networks in P-space poses significant challenges. In this study, we present an empirical analysis of the distribution of the number of a line’s nodes and the properties of the networks generated by these data in P-space. To gain insights into the operational mechanisms of the network of these data, we propose an event–link model that incorporates new nodes and links in P-space based on actual data characteristics using real data from marine and public transportation networks. The entire network consists of a series of events that consist of many nodes, and all nodes in an event are connected in the P-space. We conduct simulation experiments to explore the model’s topological features under different parameter conditions, demonstrating that the simulation outcomes are consistent with the theoretical analysis of the model. This model exhibits small-world characteristics, scale-free behavior, and a high clustering coefficient. The event–link model, with its adjustable parameters, effectively generates networks with stable structures that closely resemble the statistical characteristics of real-world networks that share similar growth mechanisms. Moreover, the network’s growth and evolution can be flexibly adjusted by modifying the model parameters.

Figures