Modelling the impact of liner shipping network perturbations on container cargo routing: Southeast Asia to Europe application
Introduction
While the effects of market cycles (Stopford, 2009) on the overall stability of liner shipping networks have been the subject of extensive research over the years, what is less known is how the overall liner shipping transport system can be affected by perturbations to the established network topology caused by events such as infrastructure developments, natural disasters or armed conflicts. These perturbations are important because they could significantly alter transportation capabilities between regions or result in accidents that can cause loss of life, injuries, economic loss or damage to the environment (Mullai and Paulsson, 2011). Therefore, understanding the impact of liner shipping perturbations on container cargo routing and their potential related accidents is crucial for decision-makers in the maritime industry who strive at being better prepare for these events.
It is unrealistic to expect remove all uncertainty related to the potential effects of the above-mentioned events. However, this uncertainty can be reduced applying quantitative frameworks that model container routing under hypothetical scenarios of network perturbations and examining historical records of accidents related to the events evaluated. Such frameworks, however, are not simple to formulate. The variety of actors and processes within modern supply chains, and the complexity of their relationships have previously led to the development of simulation-based container models whose, application have been largely compromised by their dependency on extensive and complex sets of data which are generally not available in a many cases and regions.
One of the earliest attempts to simulate maritime container flows at a global scale was the Container World project (Newton, 2008, Bell et al., 2011). This study proposed a simulation approach in which every ship, port, liner service, shipping line, truck and rail operator was represented by a separate agent. The network was built using actual port rotations published by ocean carriers. Containers were transported via each of the agents operating based on their individual set of parameters. Although the model provided a framework for global-scale container routing, it proved to be too data intensive in an competitive industry reluctant to share the data required to maintain the model (Bell et al., 2011). This limitation hampers the application of such model for scenario analysis in regions where the required data is not available.
Alternative research efforts have focused on the development of optimisation-based models that can operate with simpler datasets yet are capable of delivering reliable results using computationally efficient mathematical programs. The network used in these models can be built from published ocean carrier schedules (Zurheide and Fischer, 2012) or computed from a liner shipping network design problem (Agarwal and Ergun, 2008). The objectives in the optimisation-based literature range between minimisation of routing costs (Wang et al., 2013), minimisation of sailing time (Bell et al., 2011), maximisation of profit from an ocean carrier point of view (Ting and Tzeng, 2004) and maximisation of volumes transported (Song et al., 2005). Tran and Haasis (2013) provide a comprehensive review of previous optimisation-based works including additional relevant features such as empty container repositioning, deterministic or stochastic shipping demand, and container routing problems in time extended networks.
This study seeks to contribute to the application of optimisation-based models for the analysis of liner shipping cargo flows affected by network perturbations, building upon earlier work by Bell et al. (2013) on cost-based container assignment. The proposed application of this model minimises expected container routing costs in order to assess changes in container cargo flows under scenarios of seismic and conflict hazards affecting the Southeast Asia to Europe trades. We examine previous studies of past similar disruptions in order to discuss their potential related accidents and network parameters affected in the aftermath of the disruption scenarios presented.
The cost-based assignment model has a series of features that make it suitable for the requirements of this study: First, the cost dimension is used to model the distribution of flows and aggregates a range of dependencies such as container handling and rental cost, cargo depreciation, and transit time. As such it can be used to model possible variations in costs and times that occur on the aftermath of port disruptions. Second, it includes both port capacity and link capacity constraints that can capture disrupted operational parameters in liner shipping networks. Third, the model uses a virtual network approach (Jourquin et al., 2008) which provides an accurate representation of liner shipping operations and allows to skip disrupted ports within an established port call sequence. Finally, we extend previous formulations adding a decision variable and penalty costs for cargo not transported allowing feasible solutions in cases where disruptions decrease network routing capacity below transport demands.
The remainder of this paper is structured as follows: Section 2 describes the methodology used by first proposing a classification scheme of network perturbations, differentiating between systemic and external. This section describes the cost-based assignment model that forms the core of the perturbation analysis framework. Section 3 provides a case-study focusing on the Southeast Asia to Europe trade where the model is applied in two scenarios of port disruption: seismic hazards and political conflicts. Lastly, Section 4 presents conclusions and outlines future work.
Section snippets
Classification scheme for liner shipping network perturbations
We define “network perturbation” as any change, positive or negative, to the existing state of main components of liner shipping networks. These include ports (nodes), routes operated by container liner services (links), vessels size (capacity), and transport demands (origin-destination pairs). Whether a perturbation is positive or negative often depends on the point of view of each stake holder. For example, the 1995 Port of Kobe disruption caused by an earthquake diverted local cargo to the
Problem instance
The model proposed in Section 2.2 is applied to a representative combined network of five liner services shown in Figs. Fig. 3.2 and Fig. 3.1. The network was constructed based on existing liner services sourced from the Port Operations Research and Technology Centre (PORTeC) Delos Database2 to assess the vulnerability of ports in Southeast Asia, namely Singapore, Port Kelang (Malaysia), Jakarta/Tanjung Priok and
Conclusions and future work
The main contribution of this paper is the application of a cost-based container assignment methodology for assessing the vulnerability of a multi-port system against natural and man-made disruptions. Changes to route and port capacity parameters allow to capture potential effects to the network on the aftermath of port disruptions while a penalty cost extension to previous model formulations allows feasible solutions even when capacity is diminished below transport demands. The virtual network
Acknowledgments
The authors extend their gratitude to the two anonymous reviewers and to colleagues at the Port Operations Research & Technology Centre (PORTeC) of Imperial College London for their valuable feedback to this work. The corresponding author would like to thank the National Secretariat of Science Technology and Innovation of the Republic of Panama for the financial support received through research fellowship No. 2199-35-2012.
References (36)
- et al.
A frequency-based maritime container assignment model
Transp. Res. Part B Methodol.
(2011) - et al.
A cost-based maritime container assignment model
Transp. Res. Part B Methodol.
(2013) - et al.
The vessel schedule recovery problem (VSRP)—a MIP model for handling disruptions in liner shipping
Eur. J. Oper. Res.
(2013) - et al.
Port safety and the container revolution: a statistical study on human factor and occupational accidents over the long period Saf. Sci.
- et al.
Dynamics of liner shipping network and port connectivity in supply chain systems: analysis on East Asia
J. Transp. Geogr. Spec. Sect. Altern. Travel Futur.
(2011) Patterns of maritime supply chains: slot capacity analysis
J. Transp. Geogr.
(2011)- et al.
Managing port-related supply chain disruptions: a conceptual paper
Asian J. Shipp. Logist.
(2014) - et al.
A policy making framework for resilient port infrastructure systems
Mar. Policy
(2010) - et al.
A grounded theory model for analysis of marine accidents
Accid. Anal. Prev.
(2011) - et al.
Container routing in liner shipping
Transp. Res. Part E Logist. Transp. Rev.
(2013)
Ship scheduling and network design for cargo routing in liner shipping
Transp. Sci.
Impact of the recent earthquake and tsunami on Chilean port
Saf. Reliab. Methodol. Appl.
Rescheduling ships and cargo in liner shipping in the event of disruptions
Liner Shipp.
The influence of tsunamis on moored ships and ports
J. Korean Navig. port Res.
Impact of port disruption on supply chains: a petri net approach
Cited by (19)
Natural disasters and MNE internalization: Reoptimizing subsidiary governance
2023, Journal of World BusinessHazardous material transportation problems: A comprehensive overview of models and solution approaches
2022, European Journal of Operational ResearchCitation Excerpt :The links applied for loaded-leg may be different from those for ballast-leg, carrying just the bunker fuel. Moreover, in some cases, the renting costs for vessels and the waiting cost in origin and destination ports due to scheduling decisions should be of concern (Achurra-Gonzalez et al., 2016). Assessing the transportation risk in maritime modes of transportation requires involving the vessel accidents causing a leak for liquid hazmat (e.g., oil), which leads to high economic and environmental damages.
A literature review on quantitative models for supply chain risk management: Can they be applied to pandemic disruptions?
2022, Computers and Industrial EngineeringHolistic risk assessment of container shipping service based on Bayesian Network Modelling
2022, Reliability Engineering and System SafetyEnhancing resilience through port coalitions in maritime freight networks
2022, Transportation Research Part A: Policy and PracticeCitation Excerpt :In the model, a leg is executed as a task, e.g., loading containers at one port and unloading containers at a second port. This definition follows from earlier works (Bell et al., 2013; Achurra-Gonzalez et al., 2019). The nonnegativity of demand and leg flow variables is maintained through constraints (10) and (11), where flows also cannot exceed demand.