An efficient adaptive neuro fuzzy inference system for product demand forecasting

A very important early stage and can affect other stage in manufacturing supply chain management is product demand forecasting. The forecasting result will be used in the next stage that is called aggregate production planning which will determine the production size of each product. In this study, the authors use Adaptive Neuro Fuzzy Inference System (ANFIS) to forecast monthly product demand by consumer for the next year. ANFIS that was developed by incorporating neural networks and fuzzy logic is used because it is considered capable of acquiring knowledge from data that have uncertain pattern such as consumer demand. Determination of part of historical data as system input, fuzzy membership function, and set of fuzzy rules are carefully designed for ANFIS to produce accurate results. Computational experiments show that the ANFIS produce forecasting result that close to the actual data pattern.