Please use this identifier to cite or link to this item:
Title: Development of an intelligent system for quality assurance in the garment industry
Authors: Lee, Kar Hang Carmen
Advisors: Choy, Tommy K. L. (ISE)
Ho, George T. S. (ISE)
Law, Kris M. Y. (ISE)
Keywords: Textile industry -- Quality control
Industrial efficiency
Issue Date: 2015
Publisher: The Hong Kong Polytechnic University
Abstract: Garment manufacturing is a traditional industry with a global outlook. In general, better product quality is used as a product differentiation strategy in the marketplace. However, due to the error-prone nature of the garment manufacturing processes, it is challenging to assure that the garment products are of good quality. In addition, garment products are ephemeral goods which may no longer be attractive to consumers if they fail to arrive at the retail stores before the fashion trends change. Therefore, garment manufacturers are expected not only to assure the quality of their products, but also to enhance the production efficiency. Furthermore, with the pressure brought by the rapidly changing fashion trends in recent decades, traditional inspection-oriented quality assurance (QA) strategies are no longer sophisticated enough for the garment industry to cope with the challenges posed by this new generation. Considering how process parameters are set in production is a critical factor affecting both the product quality and production efficiency, this research paves the way for a novel approach to analyze the quality problems at the parameter level. An intelligent system, termed the Fuzzy Rule-based Recursive Mining System (FRRMS), is developed for determining optimal process parameter settings which lead to improvement in product quality. The system is equipped with a process mining feature to recursively discover the hidden relationships between production process parameters and the resultant product quality. There are three modules that constitute the system: the Fuzzy Association Rule Mining Module (FARMM), the Slippery Genetic Algorithm-based Optimization Module (sGAOM), and the Decision Support Module (DSM). The FARMM collects data from the production shop floor and represents the relationships between the process parameters and product quality features in terms of fuzzy association rules. The sGAOM optimizes the fuzzy association rules obtained with the use of a novel algorithm, namely the slippery genetic algorithm in fuzzy association rule mining (sGA-FARM). The major feature of the sGA-FARM is that it produces variable-length chromosomes by imitating and transcribing the biological slippage commonly found in DNA replication. Based on the optimized rules, the DSM estimates the resultant product features when a set of process parameters are given. The output of the DSM allows the garment manufacturers to determine the appropriate process parameter settings in order to achieve the desired product features. The feasibility of the proposed system has been validated by means of case studies. By following a generic methodology designed for the development and implementation of the system, the system was implemented in a garment manufacturing company. Through a pilot run of the system in the case company, improvement in production efficiency and product quality was observed. The major contribution of this research is in the design and implementation of an intelligent system that facilitates appropriate decision making in the formulation of effective QA strategies, addressing the current needs of the garment industry. Furthermore, the deliverables of this research not only provide a means for developing the FRRMS, but also include the use of the novel nature-inspired sGA-FARM for industrial process parameter optimization, in general.
Description: PolyU Library Call No.: [THS] LG51 .H577P ISE 2015 Lee
xvii, 237 pages :color illustrations
Rights: All rights reserved.
Appears in Collections:Thesis

Files in This Item:
File Description SizeFormat 
b28245994_link.htmFor PolyU Users203 BHTMLView/Open
b28245994_ir.pdfFor All Users (Non-printable)4.82 MBAdobe PDFView/Open
Show full item record
PIRA download icon_1.1View/Download Contents

Page view(s)

Last Week
Last month
Citations as of Oct 14, 2018


Citations as of Oct 14, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.