Research on Neural Network of Enterprise Raw material production and Transportation based on Multi-objective programming

Abstract

This paper mainly studies how to make an order plan and arrange a transfer plan when the future supply capacity of the supplier is uncertain. On the basis of reasonably using the neural network algorithm to predict the future supplier supply capacity, a single objective optimization model based on 0-1 programming is established. First of all, this paper selects ten representative indicators from four aspects: the supply capacity of suppliers, the demand intensity of enterprises for materials, the supply sustainability of suppliers and the relationship between the company and suppliers, and carries on the principal component analysis. 402 suppliers were ranked. Then, based on the neural network algorithm, the maximum supply capacity of the supplier in the next 24 weeks is predicted according to the supplier's supply data in the past 5 years, and the regression accuracy is more than 90%. A 0-1 programming model aiming at the minimum weekly supply is established, with the weekly production capacity as the lower limit and the supplier's maximum supply capacity as the upper limit. The minimum number of suppliers per week is as follows: at least 44 suppliers are selected to supply raw materials. Considering the distribution of the specific order number of each supplier, the ratio method is used to determine the weight of the supplier, so as to define the weight production function, and order according to the way of maximizing the weight production function.