Toward managing demand variability by neuro-fuzzy approach

Because of globalization, fast changes of technology and short life cycle of products, enhancing the accuracy of demand forecasts becomes one of the important issues for managers. The objective of this paper is to analyze and explore given data of orders using adaptive neuro-fuzzy inference system (ANFIS) and to draw up, by ANFIS learning mechanism, the relational rules from historical order data, whereby to construct the needed forecasting model, hoping to make accurate forecasts according to the demand variability. Afterward the proposed forecasting model is compared with the conventional regression analysis and back-propagation network to verify its feasibility and validity.