Optimization Research on Vegetable Sales in Supermarkets Based on Greedy Algorithm and ARIMA Time Series Model

With the rapid development of digitalization, it is particularly important to provide superstores with accurate information on market demand and trends, and to help them develop reasonable replenishment plans and pricing strategies. This paper aims to improve the competitiveness and operational efficiency of superstores by establishing various mathematical models. For the vegetable merchandising problem, firstly, data cleaning, analysis and visualization are carried out to analyze the relationship between the sales of six categories of vegetables using Pearson's correlation coefficient and construct a multiple linear regression model to describe the sales volume; secondly, considering the cyclical relationship between the vegetable sales volume and the time, the sales volume and the wholesale price of the coming week are predicted by the ARIMA time-series model, and a mathematical planning model; finally, the greedy algorithm was applied to solve the model to derive the daily replenishment and pricing strategy for the coming week.