The Comparison of Neural Networks’ Structures for Forecasting

International Journal of Supply and Operations Management

Ilham Slimani¹  Said Achchab¹  Ilhame El Farissi² 

[1] Al-Qualsadi Research and Development Team, National Higher School for Computer Science and System analysis (ENSIAS), Mohammed V University, Rabat, Morocco;Laboratory LSE2I, National School of Applied Sciences (ENSAO), Mohammed first University, Oujda, Morocco;

关键词: neural networks;    artificial intelligence;    supply chain management;    information sharing;    demand forecasting;    game theory;   

DOI  :  10.22034/2017.2.01

来源: DOAJ

【 摘 要 】

This paper considers the application of neural networks to demand forecasting in a simple supply chain composed of a single retailer and his supplier with a game theoretic approach. This work analyses the problem from the supplier’s point of view and the employed dataset in our experimentation is provided from a recognized supermarket in Morocco. Various attempts were made in order to optimize the total network error and the findings indicate that different neural net structures can be used to forecast demand such as Adaline, Multi-Layer Perceptron (MLP), or Radial Basis Function (RBF) Network. However, the most adequate one with optimal error is the MLP architecture.

【 授权许可】

Unknown