Identification of critical sets of ports in cascading failures of global liner shipping network

Cascading failures of port congestion propagation have become increasingly frequent in the global liner shipping network (GLSN), which harms international trade. To build resilience in maritime logistics, it is vital to proactively identify critical port sets whose initial failure would trigger severe consequences of port congestion propagation. This paper proposes a methodology for critical port sets identification in the cascading failures of GLSN, which integrates a cascading failure simulation model with a meta-heuristic optimization procedure. Specifically, a cascading failure model that considers liner shipping service routes’ port rotation adjustment behavior is introduced first. Then, the identification of critical port sets is formulated as a combinatorial optimization problem, seeking sets of ports whose simultaneous initial disruptions would trigger the largest cascade size; and an improved multi-population genetic algorithm is developed to solve this optimization problem. We demonstrate the usefulness of the proposed methodology by applying it to an empirical case of the GLSN, identifying the top 10 most critial sets of ports under different given values of set size. Extensive computational experiments validate the algorithm’s effectiveness in delivering high-quality solutions of approximate optimality. Results show that the critical sets of ports and their triggered cascade sizes are largely influenced by the service route behaviors of port rotation adjustments, and the influence becomes more profound with the increase in set size. Our proposed methodology serves as a valuable tool for proactively identifying worst-case scenarios of global port congestion propagation, thereby assisting stakeholders in resilient maritime logistics management.