Automatic Pricing and Replenishment Strategy for Vegetable Products Based on XGBoost and LSTM Prediction Models

Abstract

Due to the short shelf life and imbalance between supply and demand of vegetable commodities, it is important to develop reasonable replenishment and pricing strategies. This paper takes the measured data of a superstore as an example, and develops a reasonable decision-making program by studying the relationship between the information of each commodity in the vegetable category, the breakdown of the flow, the wholesale price and the recent attrition rate. First, the distribution pattern and interrelationship of the sales volume of each category and single product of vegetables are studied, and the distribution, difference and correlation between them are analyzed. Then, XGBoost is used to predict the sales volume and then LSTM is used to predict the wholesale price, to get the relationship between the total sales volume of vegetable categories and the cost-plus pricing, and for the total daily replenishment and pricing strategy in the coming week, a planning model is set up to solve the problem by using genetic algorithm. Finally, in order to maximize the revenue of the superstore, a goal planning model is established to take the sales unit price and sales volume as variables, and mathematical calculation methods are used to determine the daily sales volume and sales unit price and the maximum revenue, so as to determine the replenishment volume of a single product and the optimal pricing strategy. The proposed model fits the actual needs, is practical, efficient in calculation, and can effectively solve the proposed problem.