Penerapan Syaraf Tiruan Untuk Memprediksi Persediaan Barang Menerapkan Algoritma Elman Recurrent Neural Network

Abstract

Inventory management is an operational issue frequently faced by minimarkets. If the inventory level is too low and demand cannot be met due to insufficient stock, it will lead to customer disappointment, and there is a possibility that customers may not return. Similarly, if the inventory level is too high, it will result in losses for the minimarket, as they need to allocate more space, face potential depreciation of the value of goods, and incur additional costs related to inventory, such as maintenance and accounting expenses. The inventory problem at PT. Indomarco Prismatama arises from the current system used, which is based on the daily sales of each store. The daily sales data is sent to PT. Indomarco Prismatama through a link every day. As a result, the goods sent by the warehouse are based on the daily sales data provided. This needs to be improved to maximize the warehouse's performance in providing goods, especially items that need to be sent to each store with varying buying patterns specific to each store. To address the inventory management issues, a robust system is required to optimize the inventory at the warehouse. The proposed system is expected to accurately calculate the inventory at PT. Indomarco Prismatama by utilizing an artificial neural network employing the Elman Recurrent Neural Network (ERNN) algorithm.