Present a mixed approach of neural network and bat algorithm to predict customer demand in the supply chain to reduce the Bullwhip effect

Many studies have addressed supply chain that shows the importance of the subject and the competition in the supply chain consisting of several companies. The previous studies address the issue of reducing the Bullwhip Effect in the supply chain, which is possible by predicting the correct amount of customer demand. This paper improves the accuracy of the model prediction and reduces the existing error in the previous models to attain an accurate and very close-to-reality forecast and also to reduce the Bullwhip Effect in the supply chain. Literature shows the absence of research on the presentation of a metaheuristic algorithm consisting of a neural network and bat algorithm to forecast supply chain demand in manufacturing companies; therefore, this article is innovative. On the other hand, no researchers have addressed Bullwhip Effect reduction using mixed metaheuristic algorithms. Therefore, this article improves the previous models, reduces the number of errors in demand forecasting, and reduces the Bullwhip Effect. For this purpose, the scalable gradient algorithm method is used for better network training. The results indicate the optimal performance of neural network training with a comparable gradient and bat algorithm on reducing the Bullwhip Effect.