© Color. Technol., 121 (2005) 53 Web ref: 20050112 Coloration Technology Society of Dyers and Colourists A knowledge-based expert system for dyeing of cotton. Part 1: Design and development T Hussain, a R H Wardman a and R Shamey b, * a School of Textile and Design, Heriot-Watt University, Netherdale, Galashiels TD1 3HF, Scotland, UK b Textile Engineering, Chemistry & Science Department, College of Textiles, North Carolina State University, Raleigh, NC 27695-8301, USA Email: rshamey@tx.ncsu.edu Received: 15 June 2004; Accepted: 7 December 2004 Artificial intelligence-based computer programs, called expert systems, have received a great deal of attention and have been used to solve an impressive array of problems in a variety of fields. Diagnosis was one of the first subjects to be investigated after the availability of digital computers, with the advent of artificial intelligence as well as the expert system technology. A knowledge-based expert system for diagnosing problems in the dyeing of cotton has been designed and developed. The performance of the system has been tested and evaluated by human experts and is deemed to be highly satisfactory. This provides a starting point for further improvements in the system. Introduction Textile dyeing is a very complex process and troubleshooting problems in dyeing is an even more complex and challenging task. A number of troubleshooting studies have established just how difficult a task it is due to the numerous variables involved and the sensitivity to minor variations in dye application, etc. [1–5]. The troubleshooting process has been compared with the criminal investigation process, the troubleshooter being a detective, post-mortem, fixing time and place, and interviewing persons in the vicinity, etc., forming a necessary part of the investigation [6]. By experience, a troubleshooter can reduce the number of possible causes to a likely few, but the confirmation of the exact cause is difficult at best. A best estimate, possible through a process of elimination, requires answers to a series of questions and/or actual laboratory tests [7]. Although some of the defects can be analysed by the practical dyer, in many cases, the defects can be analysed only by a special textile laboratory set up for this purpose, with qualified personnel and special equipment [8]. A satisfactory diagnosis entails a well-equipped testing laboratory, extensive experience in testing [9], a good knowledge of all the textile processing stages, and the interaction between the process variables and the structural features which determine the properties of the material [10], as well as a knack of problem solving. The process of troubleshooting, in textile dyeing, has been traditionally carried out by human experts/specialists and the use of computerised systems exhibiting artificial intelligence is a relatively new technology in an attempt to provide a systematic as well as efficient diagnosis process. The aim of this research was to develop a knowledge-based expert system in the knowledge domain of troubleshooting problems in the dyeing of cotton. The emphasis of this work is to provide a broad view of troubleshooting by capturing the knowledge of human experts as well as utilising the knowledge published in the literature by using a modular approach. Potential users of the system include novice textile engineers, dyeing trainees and personnel with some textile qualification but little practical dyeing experience who are hoped to perform, with the help of the expert system, at a level comparable to expert dyers. The advantage of a modular approach is that a large problem can be broken down into smaller and manageable sub-problems/modules. Through modification of existing modules or addition of new modules, the expert system can be conveniently expanded in the future to cover the latest research findings and practices. Expert system development methodology The development of knowledge-based expert systems is a more complex variant of traditional software development [11]. Different expert system methodologies and models were reviewed to select or adapt the most suitable for the project [12–18]. However, all the reviewed methodologies and models had been proposed from the point of view of artificial intelligence (AI) experts who would engineer the expert system but were not necessarily experts in the domain about which the expert system would be developed. There are two groups of people other than AI experts who might be interested in developing expert systems on their own. One of these might be the domain expert who would have been working in the domain for a long period of time and has sufficient expertise about the domain. The other might be a person from research or academia who has a good understanding of the domain, sufficient knowledge and/or access to different resources of knowledge about the domain. Obviously, these two groups of people would not have, initially, sufficient knowledge of the expert system technology but with the availability of easy-to-learn expert system tools/shells, their aspiration to develop an expert system on their own must not be over-appreciated. In this backdrop as well as the initial step for the expert system