Supermarket decision-making model based on improved genetic algorithm and XGBoost

This paper constructs a supermarket decision-making model based on improved genetic algorithm and XGBoost. It analyzes the data distribution through the Shapiro–Wilk normality test, uses the exponential function to describe the relationship between the total sales volume and the cost-plus pricing, and builds the XGBoost prediction model to optimize the replenishment cost. The profit model is improved based on the genetic algorithm, and the probability density function is introduced to correct the fitting error of sparse data. The fitness function is defined to determine the replenishment quantity and pricing strategy of single products. The data-driven model is improved by using linear regression and canonical correlation analysis. Finally, the calculation formula of typical variables is obtained, and the optimal estimation of single product replenishment and pricing is realized.