“…Trimf and trapmf are the most popular membership functions due to simplicity and computational efficiency [43]. They can perform acceptably without the need for further tuning [44] and they apply a robust way to approach the convex function [45]. The membership functions of the inputs are divided into three linguistic values (poor, satisfactory and good), while the membership functions of the outputs (QS1, QS2, QS3 and TQI) are divided into nine linguistic values (poor–excellent).…”