Voyage Optimization Algorithm for Intelligent Shipping – Considering Energy Efficiency and Collision Avoidance

Doctoral thesis, 2025

The environmental impacts of shipping pose significant challenges. The voyage optimization system is an important tool to address challenges, with optimization algorithms serving as the core of this system. The main objectives of this thesis are to develop voyage optimization algorithms to improve energy efficiency in weather routing and investigate the capability of voyage optimization algorithms for ship collision avoidance.

Voyage optimization systems rely on ship performance models to estimate energy costs and optimization algorithms to find optimal voyages. However, ship performance models of different levels of complexity may contain large uncertainties in estimating a ship’s energy consumption and emissions, while optimization algorithms should also consider other uncertain and dynamic factors for en route planning, such as weather conditions and market fluctuations, to ensure optimal operations. In this thesis, a literature study of the state-of-the-art was first conducted to review the pros/cons of the latest optimization algorithms for weather routing. Then, five strategies to improve Isochrone algorithms were investigated to propose an Isochrone-based predictive optimization (IPO) algorithm, which can achieve real-time voyage planning providing on-time arrival and energy efficiency. In addition, a learning-based EA algorithm (L-MOEA) was proposed to explore the application of ML algorithms and address weather uncertainties in weather routing systems. Furthermore, the impact of uncertain ship performance models was researched by developing a stochastic fuel consumption model for Specific Fuel Oil Consumption (SFOC) to study its influence on voyage optimization results. Finally, the developed IPO algorithm was further developed with enhanced terrain adaptation and traffic rule compliance, and the algorithm was demonstrated in confined waterways with dynamic traffic and shown to efficiently assist a ship’s real-time voyage planning for collision avoidance to ensure ship safety.

Based on the case studies applied for the developed voyage optimization algorithms in this thesis, it can be concluded that the proposed Isochrone-based predictive optimization algorithm can achieve an average of 5% energy savings by using the benefits of combined ‘Isochrone-A*’ algorithms and machine learning models. However, for weather routing applications, the impact and uncertainty due to ship energy performance models should be carefully considered to ensure reliable voyage planning and decision support. In addition, the proposed IPO-based collision avoidance algorithm can effectively optimize a ship’s voyage planning in real-time to ensure navigation safety and on-time arrival, complying with COLREGs in both confined waterways and open waters.