Modeling and analysis of packing properties through a fuzzy inference system

Erginel N.

JOURNAL OF INTELLIGENT MANUFACTURING, vol.21, no.6, pp.869-874, 2010 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 21 Issue: 6
  • Publication Date: 2010
  • Doi Number: 10.1007/s10845-009-0262-1
  • Journal Name: JOURNAL OF INTELLIGENT MANUFACTURING
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.869-874
  • Keywords: Packing properties, Design and analysis of experiments, Fuzzy logic models, Adaptive neuro fuzzy inference system, Factorial design, LOGIC, CONTROLLER, ANFIS, PLANT
  • Anadolu University Affiliated: Yes

Abstract

Packing of manufactured products is important in protecting them from damage during handling and transportation. Several materials and methods are used for packing of products and the optimum level of packing materials should be determined to minimize damage to the product. Design and analysis of experiments (DOE) could be used for this. However, fuzzy logic models can be more suitable than mathematical models derived from DOE due to the error values. This is because fuzzy logic models use several functions instead of a single function. DOE and the adaptive neuro fuzzy inference system (ANFIS) modeling approaches are employed for the modeling and analysis of packing materials with the aim of delivering minimum damage. Although the root of mean square error (RMSE) values of the ANFIS model is 5.7622 x 10(-6), the RMSE value of mathematical model from DOE is 3.57457. This result shows that the ANFIS model is more successful than the DOE model for this purpose.