Development of live poultry production optimization and order forecasting system based on product

Facilities operating in the field of production of products from poultry carry out production based on orders. The presence of a certain number of pieces of product in each poultry makes it difficult to meet the requested orders. With this situation, the high storage cost and the short product lifespan make it important for the facilities that make this type of production to leave the least product in hand and fulfill the orders. The fact that the orders from the customers are mostly from the products with a low product share in poultry, and the fulfillment of this order will cause the remaining anatomical products of the facility to remain in the warehouse. The short shelf life of poultry products and the costly storage process can cause damage to the facilities. Due to these situations, the facilities look at some data such as customer reliability, quantity ordered, and make cuts on orders to ensure that orders are more balanced at an anatomical rate. In this study, it will be predicted from which customer and product type the future order quantities of the relevant facilities will come. In line with this study, it is aimed to distribute the orders of the company in the most accurate way with the help of the estimated order numbers. Thanks to the order numbers estimated before the order date, the calculation of the poultry that may be needed and the determination of the products to be met will help the company to take early action. Prediction processes are provided by LSTM (Long Short-Term Memory), an artificial neural network, and Prophet models developed by Facebook.