Global pooling outperforms per-ship models for ship fuel prediction
Comparative study of global pooling for pre-departure ship fuel consumption prediction: Advantages over per-ship and adaptation-based approaches
Machine LearningArtificial Intelligence
Summary
Predicting how much fuel a ship will use on a trip helps with planning but has been hard because models usually work well only for ships they’ve seen before. The authors compared different modeling methods and found that a single model using information about ship identity and features works better for all ships, including new ones without past data. Their global model was more accurate than individual ship models or methods that try to adapt models to new ships. This means one well-designed model can serve a whole fleet, saving effort and improving predictions.
What this means in practice
- •For fleet operation managers: Use one global model with ship specifications to predict fuel needs before departure for all ships in a fleet, including new vessels.
- •For maritime analytics providers: Offer fuel consumption prediction services that do not require per-ship historical data but rely on global models, improving onboarding of new ships.$Commercial implications: Enables commercial fuel prediction software products that serve fleets without needing vessel-specific training data.
Authors
Heungjun Im, Bongmin Cha, Agung Nugraha, Donghyun Kim, Seongphil Kang, Minwook Choi, Jeongmin Park, Jihwan Lee
Abstract
Most fuel oil consumption (FOC) models for voyage planning are per-ship models and thus generalize poorly beyond each vessel's records. Transfer and federated learning have been reported to perform better, but under different protocols that often use inputs unavailable before departure or random data splits. It therefore remains unclear which approach performs best. In this study, we compare 11 per-ship, adaptation-based, and global methods on 21 vessels. To prevent leakage, all methods receive identical inputs known or forecast before departure, excluding engine RPM and shaft power, and are tested on intact voyage legs after all training records of each vessel. Global pooling, in which one model describes each vessel by its ship identity and seven ship specifications, performs best overall in every setting examined. For in-fleet vessels, Global LightGBM reaches an overall MAPE of 8.92%, against 11.14% for Per-Ship LightGBM, and outperforms every per-ship and adaptation-based model. Global models of five architectures, including a Transformer, all outperform the per-ship and federated models. Fine-tuning a global model for each vessel, even by low-rank adaptation of the Transformer, adds no accuracy. Without any records of a vessel, the global model reaches a median MAPE of 11.20% and outperforms a per-ship model trained on that vessel's records for 14 of 21 vessels. Adding its first records to the pooled data is as accurate as vessel-specific adaptation. Ship identity and specifications are largely interchangeable for in-fleet vessels, but only the specifications support zero-shot prediction. Its error grows as a vessel's specifications differ from the fleet's. Per-ship models and vessel-specific adaptation are therefore unnecessary under fair comparison. One global model can serve a whole fleet, including newly delivered vessels with similar specifications.