Containership delay propagation risk analysis based on effective multivariate transfer entropy.

Containership delays affect arrival times of subsequent ports and various shipping routes, putting marine logistics at risk of disruption. Analyzing containership delay propagation risks and understanding the underlying propagation mechanism are critical for optimizing port operations and minimizing disruptions in container shipping networks. To address this challenge, this study applies causality analysis to the container shipping network to investigate the propagation risks of containership delays. A novel dimensionality reduction algorithm, in which the concept of effective transfer entropy is introduced, is employed to reduce errors and quantify the containership delay propagation risk causality. By estimating the effective multivariate transfer entropy, containership delay propagation risk networks are established for three regions: global region, East Asia, and Southeast Asia, where container ports serve as nodes and the connections represent the delay propagation risk relationships among ports. The quantitative analysis results show that each port affects 3 other ports on average and is affected by the delay propagation risks in 3 ports on average. The network density in Southeast Asian region is 0.36, compared with 0.14 and 0.13 in the global and East Asian regions, respectively. This implies that the propagation risks of containership delays will be more intensive in Southeast Asian region. Non-key ports are mainly affected unidirectionally by the delay propagation risks from key ports. To further comprehend containership delay propagation risks, numerical simulations are conducted using the network Susceptible Infected Recovered model. The results indicate that ports with high outdegrees exhibit higher risks of causing delay propagation and a wider range of influence. To improve the stability of ports operations, "infection rate" and "recovery rate" should be reduced and increased, respectively. This study proposes a novel method to construct shipping network and fills the gap in research on delay propagation risks in the field of shipping and provides valuable insights into the underlying mechanisms in containership delay propagation, assisting maritime managers in making informed decisions and implementing effective governance measures to reduce containership delay propagation risks. •This study offers a new method based on causality analysis for shipping network construction. •There is no clear correlation between a port's throughput and its risk level in delay propagation. •Non-key ports tend to receive delay propagation risks from key ports unidirectionally. •The degree in the containership delay propagation risk network measures delay propagation risks. [ABSTRACT FROM AUTHOR]