Nikolaj Bläser defends his PhD thesis
Nikolaj Bläser defends his PhD thesis 'Predicting Tramp Shipping Market Dynamics'
The defense is public, and everybody is welcome; the defense is scheduled for a maximum of three hours and will be held in English.
Follow the defense online via Zoom
The Doctoral School at Department of People and Technology will host a small reception afterwards.
Supervisors and assessment
Assessment committee:
- Jasmine Lam, Professor, Management and Maritime, Technical University of Denmark, Denmark
- Cagatay Iris, Associate Professor, Faculty of Humanities and Social Sciences, University of Liverpool, United Kingdom
- Jens Classen, Associate Professor, Department of People and Technology, Roskilde University, Denmark (Chair of the committee)
PhD Supervisor:
- Line Reinhardt, Associate professor, Department of People and Technology, Roskilde University, Denmark
- Hua Lu, Professor, Department of People and Technology, Roskilde University, Denmark
Co-supervisor:
- Gabriel Fuentes, Assistant Professor, Department of Business and Management Science, NHH Bergen, Norway
Resumé
The tramp shipping market is inherently volatile and complex, making it challenging for ship owners to anticipate future developments and make optimal vessel positioning decisions. In recent years, geopolitical disruptions and shifting trade patterns have further amplified market uncertainty, increasing the need for forecasting methodologies that can anticipate market developments and support strategic fleet positioning.
Existing market forecasting approaches rely on aggregate market indicators and short forecast horizons, limiting their ability to capture the underlying mechanisms through which market dynamics emerge. At the same time, the growing availability of high-resolution Automatic Identification System (AIS) data over the past decade presents new opportunities to develop bottom-up methodologies for modelling and forecasting tramp shipping markets.
This industrial PhD thesis explores how micro-level data can be leveraged to capture and forecast the underlying dynamics of tramp shipping markets. Conducted in collaboration with TORM A/S, a Danish product tanker shipping company, the thesis focuses on the product tanker segment and presents four interrelated contributions spanning AIS data processing, maritime traffic network modelling, and market forecasting.
First, the thesis addresses foundational AIS data quality challenges through the development of DAISTIN, a scalable, graph-based method for trajectory interpolation. By formulating AIS trajectory reconstruction as a shortest-path problem, DAISTIN repairs irregular signal gaps in historical AIS records and overcomes key limitations of existing interpolation methods.
Second, the thesis introduces MATNEC, an environment-adaptive framework for constructing a maritime traffic network from AIS data. Through the proposal of advanced node clustering and edge mapping methods, MATNEC enables realistic port-to-port route generation and accurate voyage distance estimation, producing more realistic trajectories than previous methods.
Third, the thesis introduces a framework for predicting shipping demand by formulating trade flow estimation as a minimum-cost flow problem. By integrating historical trade relationships, transport costs, and realistic voyage geometries on a time-expanded network, the method enables multi-month, commodity-specific tonne-mile forecasting and supports the evaluation of "what-if" scenarios, such as geopolitical shocks or canal blockages.
Fourth, the thesis presents a simulation model for forecasting regional vessel supply over a multi-month horizon. The model represents individual vessels as autonomous agents whose destination choices and waiting times respond dynamically to regional market imbalances.
This framework forecasts the regional distribution of available tonnage with improved accuracy relative to a baseline model.
Taken together, the thesis demonstrates how bottom-up modelling can capture key mechanisms underlying tramp shipping market dynamics and support more informed strategic decision-making.
The dissertation will be available for reading at the Roskilde University Library before the defence (on-site use). The dissertation will also be available at the defence.