Great to have you in Maritime Optima. You have been working part-time with the company since we founded it in 2018, and now you are doing your master's thesis together with Western Bulk Carriers. But you are also using ShipIntel, and the AIS data stream from Maritime Optima.
Last summer I leveraged clustering of vessel trajectory data to identify berths. These are used to define legs of a vessel’s voyage and to collect data on frequent routes serviced by different vessel segments. This summer I built a machine learning model that predicts a vessel’s destination given a partial trajectory. The results are used when reliable data is not available.
I’ve learned that it is possible to solve a lot of use cases in the shipping industry by working with available data. Additionally, it’s important to build tools to enforce data quality as this makes the data-driven results more reliable.
This year I’m writing a Master’s thesis as part of the Industrial Economics and Technology Management program at the Norwegian University of Science and Technology. The thesis is within the academic field of Operations Research. In collaboration with Western Bulk, the aim is to build a mathematical model that finds optimal cargo routes for the vessels in their fleet while maximizing the profit generated by the different cargoes. What distinguishes my model from others is that it should be able to handle future uncertain cargo demand. This complicates the model, but also makes the project more interesting to work on.
I’ve always been fascinated by the importance of the shipping industry in the world economy. Transportation is needed to connect markets of differing competitive advantages. In addition, the shipping industry offers vast arbitrage opportunities if market conditions are predicted correctly. Leveraging such arbitrage opportunities optimally has been a driving motivation pursuing the scope of the Master’s thesis.
There has been a lot of research of similar problems, especially within the academic field of routing and scheduling of tramp vessels. Models, however, often have to be tailored to specific companies. Thus, it is interesting to make adjustments to existing work to the make model more relevant for Western Bulk’s environment of operation.
One of the biggest challenges will be curating real-life cargo data that the model can use. Data on future cargo trades are often uncertain and incomplete. In addition, parameters such as freight rates are most often decided through the process of negotiation and not known in advance. Building a model that is capable of handling uncertain parameter values is another challenge.