Dynamic stock model application considering the effect of the campaign in case of demand turning into lost sales

Production and inventory management has significant effects on other management operations in terms of the results it causes. Although the broad scope of the subject causes it to be addressed in different ways, main objectives are eliminating waste, minimizing costs and ensuring customer satisfaction. Investigating the compromise solution in a supply chain is possible only with an academic perspective. In this study, the problems in production and inventory management of a manufacturer operating with a vendor-managed inventory system are considered as a inventory modeling problem under demand forecasting and total costs are compared. In the first chapter of the study, supply chain and inventory management were examined conceptually and suggestions were made for the problems in practice. In the second chapter, the inventory control models in the literature are analyzed elaborately and the differences between the models based on the demand structure are indicated. Time series analysis and artificial neural networks are explained as demand forecasting methods. In the third chapter, demand forecasts for the most important stock items, classified according to annual sales volume and determined under certain assumptions, were made using exponential smoothing methods, ARIMA models and artificial neural networks methods. The forecasted values closest to real ones were transferred as inputs to the inventory model. Since the demand was predicted and had an erratic structure, a deterministic dynamic stock control model was created. Unlike other studies in the literature, the situation in which demands that are not met on time directly turn into lost sales is addressed by taking into account the promotion effect. It was found that the total cost obtained by the proposed model was significantly lower than the total cost resulting from company policies by means of the analysis.