Interval fuzzy number-based approach for modeling an uncertain fuzzy yield learning process

Abstract

An interval fuzzy number-based approach was proposed in this study to model an uncertain yield learning process. The study aimed to overcome the limitations of present methods, wherein the lower and upper bounds of the yield are generally determined by few extreme cases, thus resulting in an unacceptable widening of the yield range. In the proposed interval fuzzy number-based approach, the range of yield was divided into two sections, namely inner and outer sections, which corresponded with the lower and upper membership functions of a fuzzy yield forecast based on interval fuzzy numbers, respectively. To fulfill different managerial objectives, in this approach, all actual values are included in the outer section, whereas most of these values fall within the inner section. To derive the values of parameters in a fuzzy yield learning model based on interval fuzzy numbers, a mixed binary nonlinear programming model was proposed and optimized. The interval fuzzy number-based approach was applied to two real-time cases for evaluating its effectiveness. According to experimental results, the performance of the proposed method was superior to that of several existing methods, particularly in terms of forecasting precision for the average range. Forecasting accuracy obtained using the interval fuzzy number-based approach was satisfactory.