|Authors||D. Zhou and A. Elmokashfi|
|Title||Overload-based cascades on multiplex networks|
|Project(s)||DOMINOS: Dissecting and Modeling Interdependencies in Communication Networks|
|Year of Publication||2017|
|Place Published||NetSci 2017 conference, Indianapolis, Indiana, USA|
Although cascading failures caused by overload in interdependent/interconnected networks have been studied in the recent years, the interplay between the extent of overlapping links (inter-similarity) and overload cascades is not well understood. This is an important issue since shared links do exist in many real-world coupled transportation systems. We propose a new model for load-based cascading failures on multiplex networks (based on previous single-layer models). This model compares different network structures, coupling schemes and overload rules. More importantly, we systematically investigate the effects of inter-similarity on the robustness of the whole system under an initial intentional attack. Surprisingly, we find that inter-similarity sometimes can have both positive and negative impacts on multiplex traffic systems. It not only brings assortativity, but also causes more overloads to the same node on both layers. This also results in continuous changes of the best overlap ratio from 1 to 0 during the transition process of the system. These results provide useful suggestions for designing robust coupled traffic systems.