Single-Input-Rule-Modules connected fuzzy inference system and its applications to failure mode and effect analysis

Two major shortcomings of the traditional fuzzy inference systems (FISs) e.g., Mamdani FIS and T -S FIS models, are (i) its difficulty to form a complete fuzzy rule base when the number of required rules is large, and (ii) the fulfillment of monotonicity property. The aims of this thesis are to anal...

Full description

Saved in:
Bibliographic Details
Main Author: Jong, Chian Haur
Format: Thesis
Language:English
Published: 2015
Subjects:
Online Access:http://ir.unimas.my/id/eprint/12068/1/Jong.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Two major shortcomings of the traditional fuzzy inference systems (FISs) e.g., Mamdani FIS and T -S FIS models, are (i) its difficulty to form a complete fuzzy rule base when the number of required rules is large, and (ii) the fulfillment of monotonicity property. The aims of this thesis are to analyze the use of a relatively new FIS, i.e., Single-Input-Rule-Modules (SIRMs) connected FIS, and a simplified theorem, as a solution for aforementioned shortcomings. Besides, a monotonicity test procedure is used to evaluate the fulfillment of monotonicity property, empiricall,J In this thesis, lhe applicability of lhe SIRMs connected FIS, its simplified theorem, and a monotonicity test procedure; to real world problems (i.e., Failure Mode and Effect Analysis (FMEA) and a water level control problems) are demonstrated. A new FMEA procedure with the SIRMs connected FIS and the simplified theorem is proposed. The simplified theorem is adopted as the governing equations to allow the SIRMs connected FIS model in FMEA to be constructed. Besides, the usefulness of the proposed FMEA procedure in agriculture i.e., edible bird nest (EBN) processing and semiconductor industry are demonstrated. The contributions of this thesis are three folds, (i) simplification of a theorem and monotonicity test for real world applications, (ii) a new FMEA procedure with the SIRMs connected FIS and (iii) case studies on BBN processing and semiconductor industry.