Demand forecasting model for inventory optimization: an example of a small enterprise

This scientific article presents a demand forecasting model developed to optimize inventory at a small enterprise. The purpose of the study is to create an algorithm that will increase the accuracy of the demand forecast. It is proposed to integrate the developed algorithm into the information base of the enterprise in order to automate calculations previously performed by managers based on past data and specified algorithms. The demand forecasting model is based on time series analysis and exponential smoothing methods and takes into account seasonality and demand trend. To control the change in sales volume, as well as to evaluate the effectiveness of the forecasting formula, it is proposed to use an exponentially weighted moving average. The result will be a formula for predicting sales for a given week. To assess the effectiveness of the proposed model, the calculated indicators are compared with the actual ones for 2022. By calculating the mean and standard deviation for actual and forecast data, it is possible to estimate the cumulative forecast for a long period and the accuracy of the forecast for a week, respectively. According to the results, the algorithm provides a better consistency of the inventory level with the sales level for the week compared to the manual calculations of the manager. This algorithm can be integrated into the 1C database to improve the inventory management process.