A systematic fuzzy modeling for scheduling of textile manufacturing system

Abstract

This paper presents a fuzzy expert system for Textile manufacturing system using fuzzy cluster analysis. The proposed approach consists of two phases. The first phase is developed with an unsupervised learning and involves a baseline design to effectively identify a prototype fuzzy system. At this phase, a cluster analysis approach is implemented. For the aim of determination of the optimal values of clustering parameters, i.e., weighting exponent (m), and the number of clusters (c), Genetic Algorithms are used. At the second phase, fine tuning process is done to adjust the parameters identified in the baseline design, subject to supervised learning. This phase is realized by using approximate reasoning module. Approximate reasoning parameters are also optimized, using GAs. Finally, the proposed approach is validated by applying it to scheduling system of a Textile industry and comparing the results with a Sugeno-type fuzzy system modeling that uses subtractive clustering in its structure identification stage. The results show that the proposed fuzzy system better represents the behaviour of the complex systems, such as Textile industries.