A K-Means Clustering Algorithm to Determine Representative Operational Profiles of a Ship Using AIS Data
Jongseo Park, Minjoo Choi
Abstract
Defining the appropriate functional requirements in the early ship design stage is important in order that costs that are caused by the over- or under-specified functional capabilities do not increase. This paper presents a K-means clustering algorithm for the determination of functional requirements. It uses automatic identification system (AIS) data from a reference ship to determine the representative operational profiles, which can support decision-makers in defining the functional requirements of ships that will be performing similar missions as those of the reference ship. In a case study, we used this method as part of a ship design project, in which the functional requirements of a battery-only electric ship are defined using AIS data from a reference ship. Results indicate that the cost can be reduced by determining the functional requirements using the proposed method.