EURO 2025 Leeds
Abstract Submission

2480. Bunker Fuel Management Optimization for Dual-Fuel Engine Vessels: A Machine Learning and Mean-Variance Optimization Approach

Invited abstract in session TA-32: Low- and zero-emission solutions for maritime operations-2, stream Maritime and Port Logistics.

Tuesday, 8:30-10:00
Room: Maurice Keyworth 1.09

Authors (first author is the speaker)

1. Qian Zhao
Alliance Manchester Business School, University of Manchester
2. Arijit De
Alliance Manchester Business School, University of Manchester
3. Richard Allmendinger
Alliance Manchester Business School, The University of Manchester

Abstract

The maritime shipping industry faces increasing pressure to reduce operational costs while adopting low-emission practices. With new International Maritime Organization regulations accelerating the adoption of Liquefied Natural Gas (LNG) propulsion and operational improvements like slow steaming, efficient bunker fuel management strategies are crucial. Bunker fuel costs, accounting for 50–60% of total operating expenses, are highly volatile, posing significant financial risks.
This study develops an optimization framework for bunker fuel management in dual-fuel engine vessels. First, machine learning models predict fuel prices for Very Low Sulphur Fuel Oil and LNG, incorporating variance and covariance measures to quantify and mitigate financial risk. Second, fuel consumption is predicted using noon report data to generate interval forecasts based on relevant factors, such as speed, main engine performance, and environmental conditions. Random Forest and Gradient Boosting Regression models achieve an R² score above 0.8. These predictions enhance fuel efficiency and support emission reduction efforts.
Finally, these forecasts are integrated into a Mixed-Integer Non-Linear Programming model, which applies mean-variance optimization to determine optimal bunkering ports, fuel quantities, and fuel types. This comprehensive approach minimizes costs and risks while facilitating the transition to sustainable maritime operations through LNG adoption and improved fuel efficiency.

Keywords

Status: accepted

Back to the list of papers