Research on superstore vegetable replenishment decision-making based on time series and PSO particle swarm algorithm

Abstract

Because vegetables are easy to wear and tear and not easy to store, supermarkets will replenish vegetables every day. In order to make reasonable replenishment and pricing decisions, this paper firstly correlates the total sales data with other data, and finds that the total sales data is negatively correlated with the average purchase price and average selling price, and positively correlated with the number of discounts, and then adopts the Feedforward Neural Network model, the Evenberg-Marquardt training algorithm, and the time series analysis prediction to predict the total daily replenishment and pricing policy of each vegetable in the coming week. Marquardt training algorithm, time series analysis and forecasting, predicting the total daily replenishment and pricing strategy of each vegetable in the coming week, and at the same time, adopting the method of dynamic planning, through the use of particle swarm algorithm of PSO to solve the total replenishment of each category and pricing strategy, and then iterated to get the total daily replenishment of each category of commodities and pricing strategy in the coming week.