Research on Vegetable Sales Strategies of Supermarkets Based on Time Series Analysis and Particle Swarm Optimization

This paper conducts a comprehensive study on the replenishment and pricing problems in the sales of vegetables in supermarkets, utilizing models such as neural networks, particle swarm algorithms, and multi-objective planning to establish a replenishment and pricing strategy based on the historical transaction information, and puts forward the suggestion of collecting more data. Firstly, the distribution pattern and correlation of sales volume of vegetable categories and individual products are analyzed, and the sales volume distribution graph is drawn through monthly data and the significance test is conducted by using Pearson's correlation coefficient and Spearman's correlation coefficient. Next, the relationship between total sales volume and cost-plus pricing is studied, and the relationship is determined through scatter plot and regression analysis, it is finally found that there is no significant relationship between total sales volume and cost-plus pricing. Next, neural network and ARIMA algorithms were used to predict the sales volume and price of the vegetable category in the coming week, and the particle swarm algorithm was applied to solve the optimization model, and the optimal solution for the revenue was obtained as $2,408.5. Finally, a multi-objective particle swarm algorithm is used to optimize the selection of saleable individual products to ensure that the market demand is met while maximizing the revenue of the superstore. In addition, this study suggests collecting more data from multiple sources to improve the accuracy and applicability of the model.