The FLR–PCFI–RBF approach for accurate and precise WIP level forecasting

Abstract

Precise and accurate prediction of future level of work in process (WIP) is an important task for factory control. To this end, a fuzzy linear regression (FLR)–partial consensus fuzzy intersection (PCFI)–radial basis function network (RBF) approach is proposed in this study. In the proposed methodology, a virtual expert committee is formed, instead of calling a number of experts in the field. For each virtual expert, a corresponding FLR equation is constructed to predict the level of WIP. Each FLR equation can be fitted by solving two equivalent nonlinear programming problems, based on the opinions of virtual experts. In order to aggregate these fuzzy WIP level forecasts, a two-step aggregation mechanism is used. First, PCFI is applied to aggregate the fuzzy forecasts into a polygon-shaped fuzzy number, in order to improve precision. Then, an RBF is constructed to defuzzify the polygon-shaped fuzzy number, and generate a representative/crisp value, resulting in improved accuracy. To evaluate the effectiveness of the proposed methodology, an actual case is discussed. According to the experimental results, the proposed methodology improved the precision and accuracy of the WIP level forecasting by 68% and 27%, respectively.