The Use Of Artificial Intelligence Models In Sales Forecasting And Supply Chain Management

Résumé: This study aims to highlight the role of sales forecasting methods in administration and supply chain management. This is done by explaining how to use forecast data in mathematical modeling for the supply chain, which is distinguished by multipurpose objectives. This can only be modeled by multi-criteria methods, where we will try to apply them to RIO, a specialized Algerian firm producing yogurt, by studying its product features in order to forecast their weekly sales by using the Box-Jenkins method and artificial neural networks. Then, we will model their supply chain using a combination of compromise programming and genetic algorithms. As shown by the application of the neural network method in the field of forcasting, it gives better and more accurate results than the application of BOX-JENKINS. Also, by comparing the results obtained from modeling using the genetic algorithm method and the Compromise programming method, we concluded that the latter gives better results than the algorithmic method. In conclusion, it can be said that if the conditions of these methods are respected, the presented study can be generalized to other similar institutions. Keywords: Box-Jenkins method, Artificial neural networks, Compromise programming, Genetic Algorithms.

Publié dans la revue: مجلة الإقتصاد الجديد