A Comparison of Artificial Neural Network and Fuzzy Linear Regression in Tire Reliability Analysis
Koosha Rafiee, Ali Azadeh and Amir-Mohammad Zohrevand
The 2010 Summer Computer Simulation Conference (SCSC 10)
Ottawa, Canada, July 11-14, 2010
The purpose of this paper is to test and compare Artificial Neural Network (ANN) and fuzzy regression to model an empirical approach to the root-cause analysis of automobile tire failure. This paper seeks to estimate survival time of different kinds of tire in experiments which is conducted in laboratory conditions based on the major characteristics of tire; such as Tire age, Wedge gauge, Interbelt gauge, End of belt #2 to buttress, Peel force, percent of carbon black. Tire life data are obtained from a laboratory test, which is developed to duplicate field failures. The models help to identify the elements of tire design affecting the probability of tire failure due to the failure mode in question. Eight Fuzzy Linear Regression models are tested. In addition, each training method of Artificial Neural Network release a different result, that leads to compare the training function based on the mean absolute percentage error (MAPE). Fuzzy Linear Regression provides better estimation than Artificial Neural Network. The major reason for applying fuzzy concept and ANN is to overcome the inter-correlation problem associated with the independent variables.
Conference Manager (V2.56.8 - Rev. 1182)