An epsilon-constraint-based exact multi-objective optimization approach for the ship schedule recovery problem in liner shipping

•A new multi-objective optimization model is developed for ship schedule recovery.•A variety of recovery options are proposed within the multi-objective framework.•An exact optimization algorithm is adopted to obtain Pareto optimal solutions.•A set of experiments are conducted for a real-life shipping route.•The conducted analyses provide important insights for shipping lines. Time management is crucial for liner shipping services. A variety of unexpected events can disrupt liner shipping schedules. A real-time port capacity analysis and rescheduling the original ship operations would be necessary to counteract the negative effects of such disruptions. Different ship schedule recovery options can be adopted in response to disruptive events (e.g., ship sailing speed adjustment, skipping of disrupted ports). However, shipping lines face conflicting decisions when selecting ship schedule recovery options. As an example, the commonly-used ship speeding-up option could effectively reduce delays during the voyage but would increase the fuel cost. Similarly, the skipping of disrupted ports may substantially decrease the associated delays but would incur additional costs associated with supply chain disruptions and misconnected cargo. Nevertheless, there is a lack of analytical methods that enable the evaluation of competing objectives in ship schedule recovery and effective multi-objective solution approaches. Therefore, this study proposes a novel multi-objective model for ship schedule recovery that aims not only to minimize the total late ship arrivals at ports but also to minimize the total profit loss due to disruptive events that may occur at sea and/or at ports. An epsilon-constraint-based exact optimization algorithm is adopted to obtain optimal Pareto Fronts. The computational experiments conducted for a real-life transit route demonstrate that the adopted exact optimization algorithm is able to generate Pareto Fronts in a timely manner. Moreover, the conducted sensitivity analyses provide interesting insights regarding the effects of different disruption types and unit fuel costs on ship schedule recovery.