Improving Ship Fleet Performance Using a Non-Parametric Model

The world merchant fleet has increased in the last decade producing an increase of fuel consumption and greenhouse gas emissions (GHGs). Thus, the concerns of ship-owners to implement alternatives to improve the fleet efficiency are growing. However, shipowners are facing barriers to implement energy efficiency technologies mainly due to reliability, financial and economic constraints as well as complexity of change. Actually several shipowners are using onboard data measurements systems that collect navigation and propulsion information of their ships. Therefore, after being sent via satellite and stored in data warehouse, these data are being made available to assess the performance of their fleets. This paper describes the use of these data to generate models in order to answer to the following questions: What is the ship with least efficiency in my fleet? What is the best strategy to improve the overall efficiency of my fleet? What is the ship that I should sell in priority? What is the influence of this maintenance policy on the performance of my fleet? The application case of this paper is based on one fleet of 13 ships containing 223 trips that gather approximately 6,844 traveling days. After the definition of the key performance indicators (KPIs), a data envelopment analysis (DEA) models is discussed. Then, a multicriterion decision analysis (MCDA) model is compared to the DEA outputs. The results suggest that this new methodology can efficiently provide a multicriteria decision framework to shipowners avoiding engineers’ subjectivity. These findings offer a new way to address efficiency and performance in ship management.