Maritime logistics is one essential part of the global supply chain, enabling the globalization of economy. It is of practical importance for improving the efficiency and reliability of maritime logistics at multiple levels. This thesis studies two operational issues in maritime logistics.

The first issue is about the concern on piracy attack, a serious security threat causing the affected shipping routes more costly and less reliable. Piracy attack occurs in many areas beyond the well-reported Somalia Pirates. By now, various strategic actions have been taken to prevent piracy attacks, such as rerouting vessels to avoid the dangerous water area, forming group transit and strengthening the navy patrols. However, these actions still are not enough to eliminate the possibility of...[ Read more ]

Maritime logistics is one essential part of the global supply chain, enabling the globalization of economy. It is of practical importance for improving the efficiency and reliability of maritime logistics at multiple levels. This thesis studies two operational issues in maritime logistics.

The first issue is about the concern on piracy attack, a serious security threat causing the affected shipping routes more costly and less reliable. Piracy attack occurs in many areas beyond the well-reported Somalia Pirates. By now, various strategic actions have been taken to prevent piracy attacks, such as rerouting vessels to avoid the dangerous water area, forming group transit and strengthening the navy patrols. However, these actions still are not enough to eliminate the possibility of piracy attacks. Therefore, it is important for a commercial vessel to be equipped with operational solutions in case of piracy attacks. In particular, choosing a direction for quickly running away is a critical real-time decision for the vessel.

This thesis starts analyzing a situation where a commercial vessel finds itself being chased by one pirate skiff. The vessel wants to make a good decision to evade the chasing. We formulate such an evading problem as a nonlinear optimal control problem. We consider different policies such as maintaining a straight direction and making turns. We start with the direct heading policy where the vessel will maintain its direction, and derive the condition under which such a policy is feasible for the vessel to be safe. Then we extend to the policy in which the vessel will make turns to evade the chasing. The feasibility condition of these policies are derived, and we develop algorithms to optimize the policies under the concept of Pareto-optimal policy.

Based on the above result, we extend our research to study the situation with multiple pirate skiffs chasing one commercial vessel. The situation will become more complicated. For example, there exists a most conservative one-turn policy against one skiff, but that is not the case when there are two skiffs. Still, we are able to show that the policies derived against one skiff can be modified for the more challenging problem.

The second issue is about planning containers transportation in feeder lines. We consider a space allocation problem for a feeder vessel to collect/ deliver containers along its route. A feeder vessel departs from a hub port, sequentially calling for a number of ports to make container collection and delivery. There are two challenges in making the decision. First, the capacity of the vessel is shared by two types of containers, laden ones collected during the route and empty ones to deliver to each port, where the collection consumes the capacity and the delivery releases the capacity. Second, the demand has also two types, some demand having reservation made in advance but subject to random cancellation, some demand coming purely from random spot market. With the commitment of fulfilling realized demand with reservation, the vessel has to decide the fulfillment level to the demand on the spot, so as to maximize the expected revenue of the whole trip. We formulate the problem as a Markov decision process, and derive a two-dimensional threshold policy for serving the demands based on the concept of discretely concavity.

The technical contribution of the thesis lies in the application of optimization. It involves two different streams of optimization, nonlinear deterministic optimization and discrete stochastic optimization, both being hard optimization problems. We are able to successfully solve these problems with useful structural results revealed.