Search

Supply chain decisions involve different entities and activities such as sourcing of material, production of goods, and logistics. Through the efficient management of these activities, firms can gain a competitive edge by beig responsive in maximizing customer value and enhance its efficiency by minimizing operating costs. Inadequate management results in poor replenishment processes which can, in turn, lead to lack of product availability in stores, lost sales, or oversupply of products, etc.. Accurate demand forecasting plays a crucial role in reducing these problems by providing precise quantities required to order and store to satisfy the customer demands, reduce inventory and production costs, reduce safety stocks and transportation, and improve replenishment processes. However, demand forecasting for products, mainly perishable goods can be challenging for firms. Many companies, such as retailers, have struggles in accurately predicting future demand. Suppliers usually have a bulk shipment process that can cause inventory differences for the retailer's desires. Furthermore, without proper information sharing between a retailer and its suppliers, mismanagement of order quantities and production can occur. Recently, these problems are tackled through the adoption of collaboration models between supplier and retailer. More precisely, companies are encouraged to integrate inventory management models to order when required. The integrated model enhances the demand planning, inventory control and replenishment processes in the supply chain. Nonetheless, the investments in such methodologies are costly for both suppliers and retailers alike. Moreover, the benefits from implementing these models are not clear in monetary terms and are not widely discussed in the literature. This research aims to improve inventory management of a retailer located in the United Arab Emirates that is facing issues with its replenishment and forecasting processes. To achieve this purpose, forecasting models such as ARIMA, exponential smoothings and machine learning models such as Artificial Neural Networks (ANN) were built and tested for the highest accuracy. The results show that forecast models ARIMA and average of ARIMA and triple exponential smoothing achieve the best accuracy, i.e., the least forecasting error value. Implementing the forecast model needs coordination and collaboration between the supply chain entities. Therefore, collaborative planning, forecasting and replenishment collaborative, CPFR, was suggested for the retailer's supply chain through a specific framework. Finally, the benefits of the collaboration framework in terms of transportation costs were displayed using discrete-event simulation. Several scenarios and parameters were tested. Adopting the CPFR collaborative framework could reduce the transportation cost for some retail items by 42% evident in the simulation. [ABSTRACT FROM AUTHOR]