Topological Analysis of Synthetic Models for Air Transportation Multilayer Networks

Abstract

Airline transportation systems can naturally be modeled as multilayer networks, each layer capturing a different airline company. Originally conceived for mimicking real-world airline transportation systems, synthetic models for airline network generation can be helpful in a variety of tasks, such as simulation and optimization of the growth of the network system, analysis of its vulnerability or strategic placement of airports. In this paper, we thoroughly investigate the behavior of existing generative models for airline multilayer networks, namely BINBALL, STARGEN, and ANGEL. To conduct our study, we used the European Air Transportation Network (EATN) and the domestic United States Airline Transportation Network (USATN) as references. Our extensive analysis of structural characteristics has revealed that ANGEL excels the two previously introduced generative models in terms of replication of the layers of the reference networks. To the best of our knowledge, this is the first study that provides a systematic comparison of generative models for airline transportation multilayer networks.