Optimizing Maritime Energy Efficiency: A Machine Learning Approach Using Deep Reinforcement Learning for EEXI and CII Compliance

The International Maritime Organization (IMO) has set stringent regulations to reduce the carbon footprint of maritime transport, using metrics such as the Energy Efficiency Existing Ship Index (EEXI) and Carbon Intensity Indicator (CII) to track progress. This study introduces a novel approach using deep reinforcement learning (DRL) to optimize energy efficiency across five types of vessels: cruise ships, car carriers, oil tankers, bulk carriers, and container ships, under six different operational scenarios, such as varying cargo loads and weather conditions. Traditional fuels, like marine gas oil (MGO) and intermediate fuel oil (IFO), challenge compliance with these standards unless engine power restrictions are applied. This approach combines DRL with alternative fuels-bio-LNG and hydrogen-to address these challenges. The DRL algorithm, which dynamically adjusts engine parameters, demonstrated substantial improvements in optimizing fuel consumption and performance. Results revealed that while using DRL, fuel efficiency increased by up to 10%, while EEXI values decreased by 8% to 15%, and CII ratings improved by 10% to 30% across different scenarios. Specifically, under heavy cargo loads, the DRL-optimized system achieved a fuel efficiency of 7.2 nmi/ton compared to 6.5 nmi/ton with traditional methods and reduced the EEXI value from 4.2 to 3.86. Additionally, the DRL approach consistently outperformed traditional optimization methods, demonstrating superior efficiency and lower emissions across all tested scenarios. This study highlights the potential of DRL in advancing maritime energy efficiency and suggests that further research could explore DRL applications to other vessel types and alternative fuels, integrating additional machine learning techniques to enhance optimization.