Pricing and Replenishment Decision System Based on LSTM

Abstract

Fresh supermarket vegetables have a very short shelf life, and the quality of vegetables will also be affected by the storage time. If most of the goods are not sold on that day, they will not be sold on the next day. From the demand side, the sales volume of vegetables is related to the supply side of vegetables from April to October. At the same time, the sales space of the supermarket is limited, which makes a reasonable sales portfolio extremely important. The commercial overpricing replenishment decision system based on LSTM analyzes the relationship between daily sales volume and cost-plus pricing for each vegetable category and single product over three years. It predicts the sales volume and wholesale price of each category for the next week, determining replenishment volume and pricing strategy based on forecasted sales volume and average loss rate. The system considers limited sales space and uses the entropy method to calculate weights for key indicators like total profit. It employs the TOPSIS method to sort items, considering forecasted wholesale prices, to select items for pricing decisions. The system aims to maximize total profit while ensuring the supermarket meets minimum display conditions. By assisting supermarkets in formulating more reasonable pricing strategies, it aims to reduce overstock, backlog, or stock shortages, ultimately improving sales volume and operational efficiency.