INTELLIGENT NEURO-FUZZY FABRIC EVALUATION SYSTEM: A NOVEL MULTI-DIMENSIONAL STOCHASTIC FUZZY SYSTEM AND A GENERATOR OF TRAINING PATTERNS FOR AN ARTIFICIAL NEURAL NETWORK

被引:0
作者
Roberto, Baeza-Serrato [1 ]
Rocio Alfonsina, Lizarraga-Morales [1 ]
Roberto Alexander, Baeza-Diaz [1 ]
机构
[1] Univ Guanajuato, Dept Multidisciplinary Studies, Guanajuato, Mexico
来源
INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING-THEORY APPLICATIONS AND PRACTICE | 2018年 / 25卷 / 02期
关键词
backpropagation; fuzzy systems; quality; quantitative characteristics; linguistic label; GENETIC ALGORITHM; EXPERT-SYSTEM; SUPPLIER; DESIGN;
D O I
暂无
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
During the production process of textile goods, the fulfillment of design specifications is critical. However, it is a fact that a customer can buy, and be comfortable with a given garment even if it does not fulfill the design specifications. This phenomenon is due to the different anthropometric measures of the potential customers. In this paper, an alternative and flexible evaluation system are proposed to categorize textile goods in a simple language such as poor, sufficient, good or excellent. Additionally, the system provides their respective quantitative quality value. Therefore, we present a novel multidimensional stochastic Fuzzy Logic System (msFLS) based on a normal probability density function to generate training patterns of quality features for an artificial neural network. The proposed approach is based on the knowledge of the expert personnel when performing an objective measurement of the main dimensions in the design specifications. The proposed approach is comprised of three modules. In the first module, a novel multi-dimensional fuzzy system is developed, in which four linguistic labels are used. The fuzzy operation, implication, and aggregation methods are applied, and the fuzzified categorization outputs are obtained. Training patterns are generated using a Gaussian distribution function, and they are used for a multilayer perceptron in the second module as a generalization mechanism. The third module is the validation of the textile quality for multiple goods where the values are defuzzified in a range of 1-10. The performance of the system was successfully validated in a real knitted textile company in the South of Guanajuato, Mexico.
引用
收藏
页码:215 / 229
页数:15
相关论文
共 37 条
[1]  
[Anonymous], 2011, INDIAN J SCI TECHNOL
[2]  
Bayat A. B., 2013, J ENG TECHNOL-US, V3, P76
[3]  
Bector C. R., 2004, FUZZY MATH PROGRAMMI
[4]  
Belliveau P., 2004, PDMA TOOL BOOK 2 NEW
[5]  
Berkan R C, 2000, FUZZY SYSTEMS DESIGN
[6]  
Bishop D.P., 1996, Text Prog, V26, P1
[7]  
Brown M., 2004, NEW PRODUCT DEV GUID
[8]   Development of a rule selection mechanism by using neuro-fuzzy methodology for structural vibration suppression [J].
Chen, Chuen-Jyh ;
Yang, Shih-Ming ;
Chen, Chu-Yun .
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS, 2013, 25 (04) :881-892
[9]  
Echeverri Jaime, 2012, Rev. ing. univ. Medellín, V11, P139
[10]   Design of a neuro-fuzzy-regression expert system to estimate cost in a flexible jobshop automated manufacturing system [J].
Fazlollahtabar, Hamed ;
Mahdavi-Amiri, Nezam .
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY, 2013, 67 (5-8) :1809-1823