Pricing and replenishment decision problem based on time series prediction and optimization algorithm

The pricing of vegetable products is an important research topic that involves multiple aspects such as agricultural production, market supply and demand, consumer interests, and social benefits. Vegetable products have characteristics such as short production cycles, short shelf life, large price fluctuations, and high demand elasticity, which pose challenges to pricing decisions. To develop replenishment decisions that maximize the profits of supermarkets, the author establishes a mathematical model for the total sales volume and cost add-on pricing of each vegetable category. First, the cost, cost add-on coefficient, and loss rate of individual products in the same category are summed according to the weight of sales, and the corresponding average cost, cost add-on coefficient, and average loss rate of the category are obtained. This initially achieves a negative correlation between the sales trend and the cost add-on coefficient using a function. Then, by constructing a LSTM neural network model and using known data to predict the sales of each category, the predicted sales volume of the model is used as the sales benchmark value. The relationship function between the sales trend and the cost add-on coefficient is supplemented, and the optimization algorithm is used to find the cost add-on coefficient value, the total daily replenishment quantity, and the maximum profit value that maximize the profits of supermarkets.