Container vessels in liner shipping are operated on closed-loop routes following a pre-announced schedule. In practice, when a vessel is sailed on the sea, there are lots of uncertain events which may delay a vessel from its original schedule, even if some uncertainty has been considered in the tactical network design. In this thesis, we investigate two issues about disruption events.

We first consider the aftermath of a disruption that delays a vessel from its given schedule, aiming to design a scheme for the vessel to catch up with the schedule in an effective way. We consider different operational actions such as speeding up, port skipping, and port swapping. For the case where only speeding up is allowed, we approach the problem by nonlinear programming and obtain certain st...[ Read more ]

Container vessels in liner shipping are operated on closed-loop routes following a pre-announced schedule. In practice, when a vessel is sailed on the sea, there are lots of uncertain events which may delay a vessel from its original schedule, even if some uncertainty has been considered in the tactical network design. In this thesis, we investigate two issues about disruption events.

We first consider the aftermath of a disruption that delays a vessel from its given schedule, aiming to design a scheme for the vessel to catch up with the schedule in an effective way. We consider different operational actions such as speeding up, port skipping, and port swapping. For the case where only speeding up is allowed, we approach the problem by nonlinear programming and obtain certain structural results of the optimal recovery schedule. It shows that speeding up can effectively handle a delay which is not too large. When there is a large delay which may be called a major disruption, we study the problem with more options such as port skipping and swapping, and develop dynamic programming algorithms on the discretized time space. We also provide a method to estimate a lower bound of the problem, which enables us to evaluate the relative error caused by the discretized time space in dynamic programming. Numerical studies are conducted to validate our results and derive managerial insights.

We then move on to consider the case where the disruption events may be partially known before its occurrence, and updating over time. For example, the weather condition or industry action at the next port that force the closure of the port can be more accurately forecasted as the vessel approaches the port. Therefore, it is essential to perform a real-time scheduling to mitigate the impact of forecasted disruption events with some uncertainties. We introduce the regular uncertainties into the problem. We propose a dynamic programming model for the real-time scheduling problem with regular uncertainties and disruption events in liner shipping, and provide the optimal control policy for the problem. Based on two kinds of terminal operations, certain structural results of the optimal operational policy are established, which represents the qualitative relationship of the optimal control actions in different delay situations. We also provide a wide range of numerical studies to verify the analytical results and demonstrate the effectiveness of the model.

In the third part, we focus on a fundamental assumption in vessel scheduling. In maritime network planning and vessel routing, it is usually assumed that each leg between two ports has a given fuel consumption function of the vessel speed. Then a speed can be derived for each leg along the route. We argue that this speed should be interpreted as an average speed of the leg, when the vessel has opportunities to save fuel by adjusting its speed several times within the leg. In this part, we aim to investigate the properties of the fuel consumption resulted by speed adjusting as a function of the average speed, with respect to monotonicity and convexity which are often desired in network planning models. We give an affirmative result to this problem, which at least partially validates the prevailing approaches used in maritime network design.