Automated Pricing and Replenishment Decision-Making for Vegetable Products Based on Statistical Optimization Models

Abstract

This article uses a statistical optimization model to address pricing and replenishment decision-making issues for supermarkets. Firstly, it visually presents the distribution patterns of sales volumes for different vegetable types and individual items. Secondly, it analyzes the sales volume distribution patterns of six categories of individual vegetables through K-means clustering. Then, it calculates the correlation between different vegetable categories and individual items by using Pearson’s correlation coefficient. Finally, it predicts the sales volume of each category for the next seven days using a time series forecasting model and determines the pricing strategy by using a cost markup method. The following conclusions are drawn: (1) Kale has the highest sales volume and significant seasonal variations; peppers and edible mushrooms have relatively large sales volumes and obvious seasonal variations. (2) The sales order of the six clustered categories is 3>5>2>4>6>1; the seasonal differences of the 2nd, 3rd, and 4th categories are significant. (3) There is a strong correlation between cauliflower and kale, a relatively strong correlation between peppers and kale, a weak correlation between eggplant and the other five vegetables, and a strong correlation between edible mushrooms and aquatic root vegetables. In summary, this article uses the aforementioned models to visualize the distribution patterns of vegetable sales volumes, determine correlations using methods such as Pearson’s correlation coefficient and time series forecasting, predict the sales volume of each category for the next seven days, and determine pricing strategies. This will help vegetable supermarkets make pricing and replenishment decisions for different vegetable categories without knowing specific items or purchase prices.