The Whiplash Effect: Congestion Dissipation and Mitigation in a Circulatory Transportation System

Abstract

The pandemic time witnessed a significant increase in port congestion, leading to shipping delays and rising costs for shippers. We build a fluid model to investigate how disruptions at one port can affect both the disrupted port and its counterpart in another country in a circulatory system where a stream of fleets transport goods back and forth between the two ports. Port disruption leads to two types of congestion: the inbound backlog, which occurs when ships are unable to enter the disrupted port, and the outbound backlog, which arises when goods are unable to be loaded onto ships for transport to other ports. We provide an analytical expression for the recovery time of the system of two ports (from when the disruption ends to when the system goes back to normal) and track the evolution of backlogs of goods and ships during the recovery process. We identify a whiplash effect in the outbound backlog level at both ports, which bears a resemblance to the commonly known "bullwhip effect". Notably, the whiplash effect manifests in three primary features, namely oscillation, attenuation, and lag. Furthermore, we extend our analysis to a network of ports and show that the key findings and insights derived from the two-port model still hold in the multi-port bipartite system. This finding confirms that, despite its parsimony, the two-port system sufficiently captures the impact of port disruptions. We also extend the fluid model to a diffusion approximation model. Finally, we apply machine learning techniques to predict the time that vessels spend in the Shanghai port and show that our proposed model reduces prediction errors compared to the benchmark, demonstrating the potential and power of our model in helping to predict and mitigate the impact of disruptions in a circulatory transportation system, e.g., those in container shipping and air travelling industries.

Keywords: supply chain disruption, risk management, port congestion, whiplash effect, machine learning