Stock planning based on time series and fuzzy forecasting methods: An application in the spring industry

In today's environment, where the concept of the market has become more localized and fragmented, being prepared for the future is of critical importance for businesses. Such preparedness is not only essential for meeting customer demands but also serves as a vital tool in competing with rivals. Businesses must analyze the market in which they operate thoroughly and develop solutions that enable them to overcome emerging problems using their own experience and capabilities. One of these problems is forecasting future demand and planning appropriate inventory levels. In this study, demand forecasting for a company operating in the spring manufacturing sector was conducted using econometric models and fuzzy time series models. Despite the existence of companies capable of producing a variety of products from various raw materials, there is a lack of studies in the spring industry that apply stock policies—based on the results of demand forecasts—to achieve cost advantages. For this reason, the study was conducted in the spring manufacturing sector. Box-Jenkins time series models were implemented using the EViews 13 software, while fuzzy time series models were applied via Microsoft Excel. Since the company has the capability to produce springs of various diameters from different raw materials, an ABC analysis was first applied among the raw materials and then among the diameters of the selected material. Once the raw material was identified, the study proceeded along two paths: demand forecasting was performed using both the ARIMA model and fuzzy time series models. Among all the models, the ARIMA(1,1,8) model yielded the lowest Mean Absolute Percentage Error. Following the application of fuzzy time series models, the method proposed by Chen (2002) was found to have the lowest forecast error based on MAPE. Forecasting was carried out using this model, and monthly forecasts were made for the following 12-month period. Based on these forecasts, the Economic Order Quantity that would minimize the company's costs was calculated. The calculated Economic Order Quantity provided a cost advantage compared to the current stock policy implemented by the company.