Designing a model for the stability of the food industry supply chain in East Azerbaijan using artificial neural network (Northwest region of the country)

The aim of research was to design a sustainable supply chain model with an artificial intelligence approach in the food industry of East Azarbaijan province in the northwest region. The statistical population included top managers, production, operations and trade of food industry companies of the province and included experts of food industry companies of province. The findings of the research showed that the resulting neural network architecture includes 4 input layers, 1 intermediate layer with 2 units and 1 output layer with two units (0 and 1). 65.8% of data ( 50 data) were used as training samples and 34.2% of data ( 26 data) were used as testing samples. The error value obtained from fitting this model is equal to 25.30 in the training sample and the error value obtained in the testing sample is equal to 7.35. Since the amount of error obtained in the testing sample is less than the traninig sample, therefore, the fit of the model was acceptable. The order of importance of independent variables in neural network structure model was estimated in the order of first to fourth priority, cultural factors, economic factors, environmental factors and social factors. The results of the research showed that the sustainable supply chain in the food industry includes local, regional, national and international arenas and creates a series of shorter and independent transactions between producers, processors, manufacturers and retailers. sustainable supply chain can cause globalization, market structure and power, consumer taste and lifestyle, change and regulation of technology.