The Robustness Of The Worldwide Marine Transport Network

The worldwide marine transport network (WMTN), the oldest transportation network, carrying goods from all over the world, is an important part of the economic globalization. While, the WMTN is also an infrastructure network, whose change may affect the global economy greatly, for instance, the earthquake and tsunami in Japan in 2011, The big fire in Tianjin seaport in 2015. The numerical simulations will be applied to the study of the robustness of the WMTN to detect its tolerance of random failure and intentional attacks.The robustness of WMTN is studied by two means, random failure and intention-al attacks. When the nodes (seaport) in the network are under random failure, the average degree, the clustering coefficient, the size of the maximum connected com-ponent and transfer efficiency of the network decreases slowly, which shows that the network is robust with respect to random fsilures; When the nodes are attacked by intentionally, the same above quantities drop rapidly, which shows that the network is fragile under attacks. Four types of intentional attacks are used, the degree cen-trality, the betweenness centrality, the eigenvector centrality and the modal analysis. Their influences on the average degree are very similar. The eigenvector centrality has a significant impact on the clustering coefficient particularly. As the node deletion ratio/= 0.28, the clustering coefficient C= 0. While, the betweenness centrality attack nearly destroys the connected component and changes the transfer efficiency most rapidly. As the node deletion ratio f is 0.15, the size of the maximal connected component S and the transfer efficiency E approaches 0.The above robustness study is also done on the network edges. Whether the edges are under random failures or intentional attacks, the clustering coefficient, the number of the isolated node and the size of the maximal connected component decreases slowly with the edge deletion ratio, which implies that the WMTN is strongly connected. The Laplacian energy decreaseS slowly with the edge deletion ratio when the edges are under random failure, while the maximum weighted edges are attacked intentionally, the Laplacian energy decreases rapidly.It can be seen from the simulation results that the WMTN can maintain its function when parts of the seaports are under random failure, but the network is easy to collapse when some important nodes are attacked intentionally. Whether the edges in the network are under random failure or intentional attacks, the connectivity changes little, but the Laplacian energy decrease rapidly when maximum weighted edges are attacked intentionally.