Affiliations: [a] School of Electrical Engineering and Computing, The University of Newcastle, Callaghan, NSW, Australia
| [b] School of Information Management, Wuhan University, Wuhan, PR China
| [c] Discipline of Business Analytics, The University of Sydney Business School, The University of Sydney, Camperdown, NSW, Australia
Correspondence:
[*]
Corresponding author. Raymond Chiong, School of Electrical Engineering and Computing, The University of Newcastle, Callaghan, NSW 2308, Australia. E-mail: [email protected]. and Zhongyi Hu, E-mail: [email protected].
Abstract: Residuary resistance prediction is an important initial step in the process of designing a sailing yacht. Being able to predict the residuary resistance accurately is crucial for calculating the required propulsive power and ensuring good performance of the sailing yacht. This paper presents a two-layer Wang-Mendel (WM) fuzzy approach to improve the approximation ability of the WM model for this prediction task. Unlike the traditional WM method, in which the consequent of its fuzzy rules is a fuzzy set, the consequent of our proposed approach corresponds to a fuzzy rule base. We apply a top-down method and fuzzy-rule clustering to construct the two-layer WM model, while a bottom-up method is employed to predict the residuary resistance. Experimental results based on two benchmark functions and a yacht hydrodynamics application show that the proposed approach is able to obtain improved robustness and accuracy in predicting residuary resistance compared to other WM model variants and well-known machine learning algorithms.
