Cum laude for Optimization of Hinterland Transportation dissertation

Electronics, clothes, food, and almost all other imported goods are transported in shipping containers. At large deep-sea ports, such as the Port of Rotterdam, tens of thousands are unloaded and loaded daily. In order to handle that many containers, it is essential that they are transported to their final destination with low costs and without taking too much time of the limited crane availability at the deep-sea terminals. The optimization of this so-called hinterland transportation is the focus of the PhD research of Bernard Zweers of CWI’s Stochastics group. He has conducted this research in a Public-Private Partnership with container terminal CTVrede in the port of Amsterdam. Zweers defended his PhD thesis ‘Optimization of hinterland container transportation and terminal operations’ at VU Amsterdam on 19 January 2021, for which he received a ‘cum laude’ degree.

Advantages of barges

In the Netherlands, the two main modes of hinterland transportation are barges and trucks. Compared with trucks, barges are cheaper and emit fewer greenhouse gasses. Moreover, they also do not contribute to more road congestion. Therefore, the European Commission and the Port of Rotterdam aim for a modal shift from trucks towards barges. Unfortunately, this modal shift has not yet taken place as desired. One of the reasons for the lack of the modal shift is that planning barge transportation is challenging. Barges are only more efficient than trucks if sufficient containers are loaded on them. Furthermore, their planning needs to be made earlier than for trucks, making it more vulnerable to disruptions. A final disadvantage of transporting containers by barge is the necessity of an extra stop at an inland terminal, because most companies do not have their own waterway connection.

Mathematical optimization models

In his dissertation, Zweers tries to overcome these challenges for barge transportation by translating practical operational problems into mathematical optimization models. By incorporating uncertainty in these models, not only the costs are taken into account, but also reliability. For instance, in one situation, this reliability is measured as the shipment’s probability of arriving on time at its destination. In another problem, a method is developed to distribute a crane driver’s workload at a terminal more evenly over the day. As a result, the workload at peak times is lower, and thus the chance of a delay is also minimized.

Efficient algorithms

Zweers presents algorithms that can find the optimal solution for the newly introduced problems. These methods work good for problems consisting of a small number of containers. However, for problems dealing with many containers the computation times can be multiple hours. Zweers shows that the problems belong to a class of problems for which it is expected that no fast method exists that can compute the optimal solution. Hence, he has also developed methods that make calculations in just a few seconds. The solutions produced by these fast methods are shown to be close to the optimal solution.

All in all, in this dissertation, problems associated with hinterland container transportation and operations at terminals are translated into mathematical models that incorporate reliability. Moreover, by presenting fast methods to solve these problems in practice, the modal shift from trucks to barges becomes easier.

Source: www.cwi.nl 

TKI Dinalog is proud to have contributed to this excellent research result which was partly conducted in the context of the TKI Logistics project Lane Analysis & Route Advisor.